Anne H Schistad Solberg

• Bhoodoo, Preetraj; Thomas, Sarina; Wetzer, Elisabeth; Solberg, Anne H Schistad & Ben-Yosef, Guy (2026). Spatio-Temporal Landmark Detection via Selective Fine-Tuning of Echocardiography Foundation Models. Proceedings of Machine Learning Research (PMLR). 307. doi: https:/openreview.net/forum?id=91igqpzV9M&noteId=91igqpzV9M. Full text in Research Archive Show summary

  Foundation models (FMs) have shown remarkable capabilities across computer vision tasks, yet their effectiveness for complex medical downstream tasks remains underexplored. This work investigates whether state–of–the–art video–based FMs for echocardiography can perform precise spatio–temporal landmark detection without extensive fine–tuning. We evaluate two recent powerful FMs, namely EchoPrime, and PanEcho, pre–trained on few millions of echocardiographic video–text pairs, for left ventricular contour detection at end–diastole (ED) and end–systole (ES) on EchoNet–Dynamic. We compare encoder regimes (frozen, partially frozen, fully trainable) and decoder heads (multilayer perceptron (MLP) vs. graph convolutional network (GCN)), and benchmark against strong non–FM backbones (ResNet–18 2D/3D, ViT–Base, MViTv2–Small). Frozen encoders perform poorly and variably (≈78.00 Dice, ED), whereas selectively unfreezing two blocks with GCN+augmentation yields a large jump (91.71 ±3.49 Dice, ED), recovering most of the improvement. Fully trainable EchoPrime (GCN+augmentation) achieves 93.13 ±3.11/90.95 ±3.71 Dice (ED/ES), which is SOTA for regression-based models on EchoNet. Deploying separate, fully fine–tuned models for each task quickly becomes impractical in resource–constrained settings. Our results suggest that partially fine-tuning the FM is a resource-efficient strategy that recovers most of the performance benefits of end-to-end training, while avoiding the overhead of maintaining a separate model for each task. The code is available at https://github.com/preetrajb/EchoVLMLandmarks.

• Mirar, Hani Nozari; Snare, Sten Roar & Solberg, Anne H Schistad (2025). Motion-Compensated Interpolation in Echocardiography: A Lie Advection-Based Approach. IEEE Transactions on Biomedical Engineering. ISSN 0018-9294. 72(1), p. 123–136. doi: 10.1109/TBME.2024.3440838. Full text in Research Archive
• Aasan, Marius; Kolbjørnsen, Odd; Solberg, Anne H Schistad & Rivera, Adín Ramírez (2025). A Spitting Image: Modular Superpixel Tokenization in Vision Transformers. Lecture Notes in Computer Science (LNCS). ISSN 0302-9743. 15641, p. 124–142. doi: 10.1007/978-3-031-93806-1_11. Full text in Research Archive Show summary

  Vision Transformer (ViT) architectures traditionally employ a grid-based approach to tokenization independent of the semantic content of an image. We propose a modular superpixel tokenization strategy which decouples tokenization and feature extraction; a shift from contemporary approaches where these are treated as an undifferentiated whole. Using on-line content-aware tokenization and scale- and shape-invariant positional embeddings, we perform experiments and ablations that contrast our approach with patch-based tokenization and randomized partitions as baselines. We show that our method significantly improves the faithfulness of attributions, gives pixel-level granularity on zero-shot unsupervised dense prediction tasks, while maintaining predictive performance in classification tasks. Our approach provides a modular tokenization framework commensurable with standard architectures, extending the space of ViTs to a larger class of semantically-rich models.

• Andreassen, Børge Solli; Thomas, Sarina; Solberg, Anne H Schistad; Samset, Eigil & Volgyes, David (2024). Introducing Anatomical Constraints in Mitral Annulus Segmentation in Transesophageal Echocardiography. In Gomez, Alberto; Khanal, Bishesh; King, Andrew & Namburete, Ana (Ed.), Simplifying Medical Ultrasound: 5th International Workshop, ASMUS 2024, Held in Conjunction with MICCAI 2024, Marrakesh, Morocco, October 6, 2024, Proceedings. Springer Nature Switzerland. ISSN 9783031736476. p. 122–131. doi: 10.1007/978-3-031-73647-6_12. Full text in Research Archive Show summary

  The morphology of the mitral annulus plays an important role in diagnosing and treating mitral valve disorders. Automated segmentation has the promise to be time-saving and improve consistency in clinical practice. In the past years, segmentation has been dominated by methods based on deep learning. Deep learning-based segmentation methods have shown good results, but their consistency and robustness are still subjects of active research. In this work, we introduce a method that combines Graph Convolutional Networks with a 3D CNN model to integrate an anatomical shape template for the predictions. Our method leverages the feature extraction capability of CNN models to provide input features to the graph neural networks. The proposed method leverages strengths from a shape model approach with the strengths of deep learning. Further, we propose loss functions for the CNN designed to guide the graph model training. The CNN was trained with transfer learning, using a limited number of labeled transesophageal echocardiography volumes to adapt to the mitral annulus segmentation task. When comparing the segmentation of the mitral annulus achieved by the proposed method with the test set annotations, the method showed a high degree of accuracy, achieving a curve-to-curve error of 2.00±0.81 mm and a relative perimeter error of 4.42±3.33%. Our results show that the proposed method is a promising new approach for introducing anatomical template structures in medical segmentation tasks.

• Hajj, Ghadi al; Hubin, Aliaksandr; Kanduri, Chakravarthi; Pavlovic, Milena; Rand, Knut Dagestad & Widrich, Michael [Show all 11 contributors for this article] (2024). Incorporating probabilistic domain knowledge into deep multiple instance learning. Proceedings of Machine Learning Research (PMLR). 235, p. 17279–17297. Full text in Research Archive Show summary

  Deep learning methods, including deep multiple instance learning methods, have been criticized for their limited ability to incorporate domain knowledge. A reason that knowledge incorporation is challenging in deep learning is that the models usually lack a mapping between their model components and the entities of the domain, making it a non-trivial task to incorporate probabilistic prior information. In this work, we show that such a mapping between domain entities and model components can be defined for a multiple instance learning setting and propose a framework DeeMILIP that encompasses multiple strategies to exploit this mapping for prior knowledge incorporation. We motivate and formalize these strategies from a probabilistic perspective. Experiments on an immune-based diagnostics case show that our proposed strategies allow to learn generalizable models even in settings with weak signals, limited dataset size, and limited compute.

• Thomas, Sarina; Tiago, Cristiana Ferreira; Andreassen, Børge Solli; Aase, Svein Arne; Šprem, Jurica & Steen, Erik [Show all 8 contributors for this article] (2023). Graph Convolutional Neural Networks for Automated Echocardiography View Recognition: A Holistic Approach. In Kainz, Bernhard; Noble, Alison; Schnabel, Julia; Khanal, Bishesh; Müller, Johanna Paula & Day, Thomas (Ed.), Simplifying Medical Ultrasound: 4th International Workshop, ASMUS 2023, Held in Conjunction with MICCAI 2023, Vancouver, BC, Canada, October 8, 2023, Proceedings. Springer Nature Switzerland. ISSN 9783031445217. p. 44–54. doi: 10.1007/978-3-031-44521-7_5. Full text in Research Archive Show summary

  To facilitate diagnosis on cardiac ultrasound (US), clinical practice has established several standard views of the heart, which serve as reference points for diagnostic measurements and define viewports from which images are acquired. Automatic view recognition involves grouping those images into classes of standard views. Although deep learning techniques have been successful in achieving this, they still struggle with fully verifying the suitability of an image for specific measurements due to factors like the correct location, pose, and potential occlusions of cardiac structures. Our approach goes beyond view classification and incorporates a 3D mesh reconstruction of the heart that enables several more downstream tasks, like segmentation and pose estimation. In this work, we explore learning 3D heart meshes via graph convolutions, using similar techniques to learn 3D meshes in natural images, such as human pose estimation. As the availability of fully annotated 3D images is limited, we generate synthetic US images from 3D meshes by training an adversarial denoising diffusion model. Experiments were conducted on synthetic and clinical cases for view recognition and structure detection. The approach yielded good performance on synthetic images and, despite being exclusively trained on synthetic data, it already showed potential when applied to clinical images. With this proof-of-concept, we aim to demonstrate the benefits of graphs to improve cardiac view recognition that can ultimately lead to better efficiency in cardiac diagnosis.

• Jahren, Tollef Struksnes; Sørnes, Anders Rasmus; Dénarié, Bastien Emmanuel; Steen, Erik; Bjåstad, Tore Grüner & Solberg, Anne H Schistad (2023). Reverberation Suppression in Echocardiography Using a Causal Convolutional Neural Network. IEEE Access. 11, p. 67922–67937. doi: 10.1109/ACCESS.2023.3292212. Full text in Research Archive
• Andreassen, Børge Solli; Volgyes, David; Samset, Eigil & Solberg, Anne H Schistad (2022). Mitral Annulus Segmentation and Anatomical Orientation Detection in TEE Images Using Periodic 3D CNN. IEEE Access. 10, p. 51472–51486. doi: 10.1109/ACCESS.2022.3174059. Full text in Research Archive Show summary

  Segmentation of the mitral annulus is often an important step in cardiac examinations. We propose a robust 3D method for predicting the anatomical orientation and segmentation of the mitral annulus in 3D transesophageal echocardiography. The method takes advantage of the circular anatomy of the annulus by utilizing cylinder coordinate samples and a 3D convolutional neural network with circular convolutions. Furthermore, the paper proposes new landmark detection loss functions based on the earth mover’s distance. The method’s effectiveness was demonstrated by training a HighRes3dNet model and evaluating its performance on a separate test set consisting of 135 frames from 19 examinations. The obtained coordinate prediction error was 1.96± 1.62 mm, and the anatomical orientation prediction error was 9.7° ± 15.8°. The robust and fully automatic mitral annulus segmentation and orientation prediction provided by the method can ease the workload of clinicians and provide time savings in clinics.

• Wajdan, Ali; Jahren, Tollef Struksnes; Villegas-Martinez, Manuel; Khan, Faraz Hameed; Halvorsen, Per Steinar & Odland, Hans Henrik [Show all 9 contributors for this article] (2022). Automatic Detection of Aortic Valve Events Using Deep Neural Networks on Cardiac Signals From Epicardially Placed Accelerometer. IEEE Journal of Biomedical and Health Informatics. ISSN 2168-2194. 26(9), p. 4450–4461. doi: 10.1109/JBHI.2022.3181148. Full text in Research Archive Show summary

  Background: Miniaturized accelerometers incorporated in pacing leads attached to the myocardium, are used to monitor cardiac function. For this purpose functional indices must be extracted from the acceleration signal. A method that automatically detects the time of aortic valve opening (AVO) and aortic valve closure (AVC) will be helpful for such extraction. We tested if deep learning can be used to detect these valve events from epicardially attached accelerometers, using high fidelity pressure measurements to establish ground truth for these valve events. Method: A deep neural network consisting of a CNN, an RNN, and a multi-head attention module was trained and tested on 130 recordings from 19 canines and 159 recordings from 27 porcines covering different interventions. Due to limited data, nested cross-validation was used to assess the accuracy of the method. Result: The correct detection rates were 98.9% and 97.1% for AVO and AVC in canines and 98.2% and 96.7% in porcines when defining a correct detection as a prediction closer than 40 ms to the ground truth. The incorrect detection rates were 0.7% and 2.3% for AVO and AVC in canines and 1.1% and 2.3% in porcines. The mean absolute error between correct detections and their ground truth was 8.4 ms and 7.2 ms for AVO and AVC in canines, and 8.9 ms and 10.1 ms in porcines. Conclusion: Deep neural networks can be used on signals from epicardially attached accelerometers for robust and accurate detection of the opening and closing of the aortic valve.

• Mirarkolaei, Hani Nozari; Snare, Sten Roar & Solberg, Anne H Schistad (2021). A robust bidirectional motion-compensated interpolation algorithm to enhance temporal resolution of 3D echocardiography. Biomedical Signal Processing and Control. ISSN 1746-8094. 65, p. 1–11. doi: 10.1016/j.bspc.2020.102384. Full text in Research Archive
• Jahren, Tollef Struksnes; Steen, Erik Normann; Aase, Svein Arne & Solberg, Anne H Schistad (2020). Estimation of End-Diastole in Cardiac Spectral Doppler Using Deep Learning. IEEE Transactions on Ultrasonics, Ferroelectrics and Frequency Control. ISSN 0885-3010. 67(12), p. 2605–2614. doi: 10.1109/TUFFC.2020.2995118. Full text in Research Archive
• Andreassen, Børge Solli; Veronesi, Federico; Gerard, Olivier; Solberg, Anne H Schistad & Samset, Eigil (2020). Mitral Annulus Segmentation Using Deep Learning in 3-D Transesophageal Echocardiography. IEEE Journal of Biomedical and Health Informatics. ISSN 2168-2194. 24(4), p. 994–1003. doi: 10.1109/JBHI.2019.2959430. Full text in Research Archive Show summary

  3D Transesophageal Echocardiography is an excellent tool for evaluating the mitral valve and is also well suited for guiding cardiac interventions. We introduce a fully automatic method for mitral annulus segmentation in 3D Transesophageal Echocardiography, which requires no manual input. One hundred eleven multi-frame 3D transesophageal echocardiography recordings were split into training, validation, and test sets. Each 3D recording was decomposed into a set of 2D planes, exploiting the symmetry around the centerline of the left ventricle. A deep 2D convolutional neural network was trained to predict the mitral annulus coordinates, and the predictions from neighboring planes were regularized by enforcing continuity around the annulus. Applying the final model and post-processing to the test set data gave a mean error of 2.0 mm - with a standard deviation of 1.9 mm. Fully automatic segmentation of the mitral annulus can alleviate the need for manual interaction in the quantification of an array of mitral annular parameters and has the potential to eliminate inter-observer variability.

• Mirarkolaei, Hani Nozari; Snare, Sten Roar; Solberg, Anne H Schistad & Steen, Erik Normann (2020). Frame rate up-conversion in cardiac ultrasound. Biomedical Signal Processing and Control. ISSN 1746-8094. 58. doi: 10.1016/j.bspc.2020.101863. Full text in Research Archive
• Waldeland, Anders U.; Coimbra, T.A.; Faccipieri, J. H.; Solberg, Anne H Schistad & Gelius, Leiv-J. (2019). Fast estimation of prestack Common Reflection Surface parameters. Geophysical Prospecting. ISSN 0016-8025. 67(5), p. 1163–1183. doi: 10.1111/1365-2478.12740. Full text in Research Archive Show summary

  We present a method for fast estimation of finite offset common reflection surface parameters. Firstly, the derivatives with respect to offset are derived from the velocity guide. Secondly, we apply structure tensors to extract the derivatives with respect to midpoint from stacked common offset sections. Finally, the mixed derivative is estimated using a one‐parametric semblance search. The proposed method is compared to the global five‐parametric semblance search and the pragmatic sequential two‐parametric semblance search on one synthetic and one real data set. The experiments show that the proposed method is more robust against noise than the pragmatic search and have comparable robustness with the global search. The proposed method smoothes parameter estimates in a local window, and the window size is set to give the best trade‐off between detail and robustness. Since the proposed method is dependent on a velocity guide, the quality of the other parameter estimates may be influenced by any inaccuracies in the guide. The main advantage of the proposed method is the computational efficiency. When compared with a gridded implementation of the semblance search, the proposed method is 10 and 400 times faster than the pragmatic and global search. Alternative search strategies significantly reduce the computational cost of the global search. However, since more than 99% of the computational cost of the proposed method comes from the semblance search to estimate the mixed derivative, it is expected that such techniques also reduce the computational cost for the proposed method.

• Waldeland, Anders Ueland; Zhao, Hao; Faccipieri, J. H.; Solberg, Anne H Schistad & Gelius, Leiv-J. (2018). Fast and robust common-reflection-surface parameter estimation. Geophysics. ISSN 0016-8033. 83(1), p. O1–O13. doi: 10.1190/GEO2017-0113.1. Full text in Research Archive Show summary

  The common-reflection-surface (CRS) method offers a stack with higher signal-to-noise ratio at the cost of a time-consuming semblance search to obtain the stacking parameters. We have developed a fast method for extracting the CRS parameters using local slope and curvature. We estimate the slope and curvature with the gradient structure tensor and quadratic structure tensor on stacked data. This is done under the assumption that a stacking velocity is already available. Our method was compared with an existing slope-based method, in which the slope is extracted from prestack data. An experiment on synthetic data shows that our method has increased robustness against noise compared with the existing method. When applied to two real data sets, our method achieves accuracy comparable with the pragmatic and full semblance searches. Our method has the advantage of being approximately two and four orders of magnitude faster than the semblance searches.

• Waldeland, Anders Ueland; Jensen, Are Charles; Gelius, Leiv-J. & Solberg, Anne H Schistad (2018). Convolutional neural networks for automated seismic interpretation. The Leading Edge. ISSN 1070-485X. 37(7), p. 529–537. doi: 10.1190/tle37070529.1. Full text in Research Archive Show summary

  Deep-learning methods have proved successful recently for solving problems in image analysis and natural language processing. One of these methods, convolutional neural networks (CNNs), is revolutionizing the field of image analysis and pushing the state of the art. CNNs consist of layers of convolutions with trainable filters. The input to the network is the raw image or seismic amplitudes, removing the need for feature/attribute engineering. During the training phase, the filter coefficients are found by iterative optimization. The network thereby learns how to compute good attributes to solve the given classification task. However, CNNs require large amounts of training data and must be carefully designed and trained to perform well. We look into the intuition behind this method and discuss considerations that must be made in order to make the method reliable. In particular, we discuss how deep learning can be used for automated seismic interpretation. As an example, we show how a CNN can be used for automatic interpretation of salt bodies.

• Kvam, Johannes; Gangsei, Lars Erik; Kongsro, Jørgen & Solberg, Anne H Schistad (2018). The use of deep learning to automate the segmentation of the skeleton from CT volumes of pigs. Translational Animal Science (TAS). 2(3), p. 324–335. doi: 10.1093/tas/txy060. Full text in Research Archive
• Breivik, Lars Hofsøy; Snare, Sten Roar; Steen, Erik Normann & Solberg, Anne H Schistad (2017). Real-time nonlocal means-based despeckling. IEEE Transactions on Ultrasonics, Ferroelectrics and Frequency Control. ISSN 0885-3010. 64(6), p. 959–977. doi: 10.1109/TUFFC.2017.2686326. Full text in Research Archive
• Landmark, Knut & Solberg, Anne H Schistad (2015). Reducing scalloping in synthetic aperture radar images using a composite image transform. Proceedings of SPIE, the International Society for Optical Engineering. ISSN 0277-786X. doi: 10.1117/12.2194952. Full text in Research Archive Show summary

  In burst mode SAR imaging, echo intensity depends on the target's azimuth position in the antenna pattern.As a result, an amplitude modulation known as scalloping may appear, particularly in ScanSAR imagesof ocean areas. A denoising method, recently developed for multibeam bathymetry, can be used to reduceresidual scalloping in ScanSAR images. The algorithm is analogous to a band-stoplter in the frequencydomain. Here, the transform is the composition of an edge detection operator and a discrete Radon transform(DRT). The edge operator accentuatesne-scale intensity changes; the DRT focuses linear features, as eachDRT component is the sum of pixel intensities along a linear graph. A descallopinglter is implemented inthe DRT domain by suppressing the range direction. The restored image is obtained by applying the inversecomposite transform. First, a rapidly converging iterative pseudo-inverse DRT is computed. The edgeoperator is a spatiallter based on a discrete approximation of the Laplace operator, but modi ed to makethe operator invertible. The method was tested on ocean scene ScanSAR images from the Envisat AdvancedSynthetic Aperture Radar. The scalloping e ect was signi cantly reduced, with no apparent distortion orsmoothing of physical features.

• Landmark, Knut; Solberg, Anne H Schistad; Albregtsen, Fritz; Austeng, Andreas & Hansen, Roy Edgar (2015). A Radon-Transform-Based Image Noise Filter - With Applications to Multibeam Bathymetry. IEEE Transactions on Geoscience and Remote Sensing. ISSN 0196-2892. doi: 10.1109/TGRS.2015.2436380. Full text in Research Archive Show summary

  This paper describes a linear-image-transform-based algorithm for reducing stripe noise, track line artifacts, and motion-induced errors in remote sensing data. Developed for multibeam bathymetry (MB), the method has also been used for removing scalloping in synthetic aperture radar images. The proposed image transform is the composition of an invertible edge detection operator and a fast discrete Radon transform (DRT) due to Götz, Druckmüller, and Brady. The inverse DRT is computed by using an iterative method and exploiting an approximate inverse algorithm due to Press. The edge operator is implemented by circular convolution with a Laplacian point spread function modified to render the operator invertible. In the transformed image, linear discontinuities appear as high-intensity spots, which may be reset to zero. In MB data, a second noise signature is linked to motion-induced errors. A Chebyshev approximation of the original image is subtracted before applying the transform, and added back to the denoised image; this is necessary to avoid boundary effects. It is possible to process data faster and suppress motion-induced noise further by filtering images in nonoverlapping blocks using a matrix representation for the inverse DRT. Processed test images from several MB data sets had less noise and distortion compared with those obtained with standard low-pass filters. Denoising also improved the accuracy in statistical classification of geomorphological type by 10-28% for two sets of invariant terrain features.

• Mirarkolaei, Hani Nozari; Snare, Sten Roar; Breivik, Lars Hofsøy; Steen, Erik Normann & Solberg, Anne H Schistad (2014). Fixed Point Motion Estimation for frame rate up conversion in cardiac ultrasound. Proceedings - IEEE Ultrasonics Symposium. ISSN 1948-5719. p. 2351–2354. doi: 10.1109/ULTSYM.2014.0586. Full text in Research Archive
• Breivik, Lars Hofsøy; Snare, Sten Roar; Mirarkolaei, Hani Nozari; Steen, Erik & Solberg, Anne H Schistad (2014). Multiscale nonlocal means method for ultrasound despeckling, 2014 IEEE International Ultrasonics Symposium Proceedings (IUS), Chicago, Illinois USA, September 3-6, 2014. IEEE (Institute of Electrical and Electronics Engineers). ISSN 9781479970490. p. 1324–1327. doi: 10.1109/ULTSYM.2014.0327. Full text in Research Archive
• Mirarkolaei, Hani Nozari; Snare, Sten Roar; Breivik, Lars Hofsøy; Steen, Erik & Solberg, Anne H Schistad (2014). Fixed Point Motion Estimation for frame rate up conversion in cardiac ultrasound, 2014 IEEE International Ultrasonics Symposium Proceedings (IUS), Chicago, Illinois USA, September 3-6, 2014. IEEE (Institute of Electrical and Electronics Engineers). ISSN 9781479970490. p. 2351–2354. doi: 10.1109/ULTSYM.2014.0586. Full text in Research Archive
• Salberg, Arnt Børre; Rudjord, Øystein & Solberg, Anne H Schistad (2014). Oil spill detection in hybrid-polarimetric SAR images. IEEE Transactions on Geoscience and Remote Sensing. ISSN 0196-2892. 52(10), p. 6521–6533. doi: 10.1109/TGRS.2013.2297193. Full text in Research Archive Show summary

  Oil spill detection in SAR images operating in a hybrid-polarimetric mode is examined. We propose and review several strategies for oil spill detection in hybrid-polarimetric SAR data. The retrieved measures are successfully applied to SAR data covering oil spill experiments outside Norway and the Deepwater Horizon incident in the Gulf of Mexico. It is shown that, under the assumption of a two-scale Bragg scattering model, a coherence measure may be recovered equally well from hybrid-polarimetric data, as for full-polarimetric data, and that this measure may be retrieved directly from the measurements without the need for any additional assumptions. The results show that low-wind lookalikes may be suppressed at the same time as the contrast of the oil spills is maintained using hybrid-polarimetric data and that multifeature images may be constructed to further enhance the oil spill detection performance. Due to the potential of wide swath widths, we conclude that hybrid-polarity is an attractive mode for future SAR-based oil spill monitoring.

• Breivik, Lars Hofsøy; Snare, Sten Roar; Mirarkolaei, Hani Nozari; Steen, Erik Normann & Solberg, Anne H Schistad (2014). Multiscale nonlocal means method for ultrasound despeckling. Proceedings - IEEE Ultrasonics Symposium. ISSN 1948-5719. p. 1324–1327. doi: 10.1109/ULTSYM.2014.0327. Full text in Research Archive
• Landmark, Knut; Solberg, Anne H Schistad; Austeng, Andreas & Hansen, Roy Edgar (2014). Bayesian seabed classification using angle-dependent backscatter data from multibeam echo sounders. IEEE Journal of Oceanic Engineering. ISSN 0364-9059. doi: 10.1109/JOE.2013.2281133. Full text in Research Archive Show summary

  Acoustical seabed classification is a technology for mapping seabed sediments. Processed multibeam sonar data yield the variation of the seabed scattering strength with incidence angle, and this paper examines the effect of this on classification. A simple Gaussian statistical model is developed for the observed scattering strength, whereby an observation is represented by a piecewise constant function of incidence angle. Provided some data for which the sediment types are known (training data), the statistics for each type can be robustly estimated. Subsequently, a standard Bayesian theory is applied to classify new observations. The model was used to compute limits on classification accuracy in terms of the intrinsic scattering strength statistics of the seabed, and to predict whether a logarithmic or linear scale for the data is preferable. Systematic experiments on a North Sea data set with four sediment classes tested how the classification accuracy depends on the piecewise function approximation, incidence angle range, amount of training data, and spatial averaging (combining consecutive pings into one observation). The classifier based on Gaussian statistics performed at least as well as sophisticated algorithms with no assumptions about the data statistics. The best accuracy (95%) was attained for logarithmic data. The amount of training data needed to achieve this was about 500 pings per class; spatial averaging could be limited to 10-20 pings. Comparable across-track spatial resolution was possible by dividing the full swath into separate independent sectors, but only at reduced accuracy (87% or less). However, comparable accuracy may be possible by taking into account the spatial relationships of observations.

• Berthelot, Angélique; Solberg, Anne H Schistad & Gelius, Leiv-J. (2013). Texture attributes for detection of salt. Journal of Applied Geophysics. ISSN 0926-9851. 88, p. 52–69. doi: 10.1016/j.jappgeo.2012.09.006. Full text in Research Archive
• Solberg, Anne H Schistad (2013). Deriving structural information from robust phase estimates. Society of Exploration Geophysicists. Expanded Abstracts with Biographies. ISSN 1052-3812. doi: 10.1190/segam2013-0985.1. Full text in Research Archive
• Solberg, Anne H Schistad; Jensen, Are Charles & Gelius, Leiv-J. (2012). Robust 2D Hilbert Atributes of Local Amplitude and Phase. In EAGE, 2012 (Eds.), 74th EAGE Conference & Exhibition incorporating SPE EUROPEC 2012. European Association of Geoscientists and Engineers (EAGE). ISSN 9789073834279. Full text in Research Archive Show summary

  Instantaneous amplitude and phase are based on the 1D Hilbert transform of seismic traces. We present alternative attributes with improved localization properties by applying a combination of orientationadaptive Hilbert transforms and bandpass filtering. The concept of local amplitude and phase are first explained for the 1D case, before extensions to 2D are given. In the 2D case, the process consists conceptually of two steps: First, the dominant local orientation is estimated, and then the local amplitude and phase are estimated along that orientation. The proposed attributes are compared to the classical instantaneous attributes both for a thin-bed resolution image with a classical wedge model, and for a seismic data set. They show better localization properties for the thin-bed wedge model, and a higher robustness and lateral continuity on the real seismic data.

• Salberg, Arnt Børre; Rudjord, Øystein & Solberg, Anne H Schistad (2012). Model based oil spill detection using polarimetric SAR, Geoscience and Remote Sensing Symposium (IGARSS), 2012 IEEE International. IEEE Press. ISSN 9781467311601. p. 5884–5887. doi: 10.1109/IGARSS.2012.6352270. Full text in Research Archive Show summary

  In this paper we propose a model-based oil spill detection approach using polarimetic SAR data. The underlying hypothesis is that the co-polarized phase difference is zero when assuming Bragg scattering mechanisms. From this hypothesis we may construct a (linear) model and derive features that discriminates oil slicks from sea water using dual-polarized (VV and HH) SAR data. We also investigate the model when assuming X-Bragg scattering, and derive a feature based on the Pauli decomposition that are invariant to the tilt angle reflection plane. The oil spill detection methodology is evaluated on a Radarsat-2 quad-pol image that covers various types of oil released in an oil-in-sea exercise in Norway. The results show that the oil spills are clearly visible in all of the derived feature-based images. Furthermore, the feature images have a more homogeneous background than the VV-polarized SAR image, in particular the Pauli-based feature image that suppresses a non-oil slick present in the VV-image.

• Landmark, Knut; Solberg, Anne H Schistad; Nilsen, Carl-Inge Colombo; Austeng, Andreas & Hansen, Roy Edgar (2012). Bayesian Seabed Classification Using Angle-Dependent Backscatter Data from Multibeam Sonar, 11th European Conference on Underwater Acoustics 2012 (ECUA 2012). Curran Associates, Inc.. ISSN 9781622761920. p. 417–424. doi: 10.1109/joe.2013.2281133. Full text in Research Archive
• Solberg, Anne H Schistad (2012). Remote Sensing of Ocean Oil-Spill Pollution. Proceedings of the IEEE. ISSN 0018-9219. 100(10), p. 2931–2945. doi: 10.1109/JPROC.2012.2196250. Full text in Research Archive Show summary

  Oil spills on the sea surface are observed relatively often. Pollution due to either accidents or deliberate oily discharges from ships represents a serious threat to the marine environment. Operational oil spill monitoring is currently done using a combination of satellite monitoring and aircraft surveillance. The combined use of satellite-based synthetic aperture radar (SAR) images and aircraft surveillance flights is a cost-effective way to monitor oil spills in large ocean areas and catch the polluters. SAR images enable covering large areas, but aircraft observations are needed to prosecute the polluter, and in certain cases to verity the oil spill. Traditionally, oil spill detection is based on single polarization SAR images. Oil spills can be discriminated from look-alikes based on a set of features describing the contrast, shape, homogeneity, source, and surroundings of the slick. Good performance is reported for single-polarization oil spill detection, but in certain cases the oil slicks cannot be discriminated from biogenic films. In the recent years, a number of studies have shown that polarimetric SAR can improve the discrimination between oil slicks and biogenic films. Several features computed from dual-pol or quad-pol images have been proposed. These include both quad-pol features like polarimetric entropy and anisotropy, mean scattering angle, polarimetric span, conformity coefficient, as well as the dual-pol features standard deviation of the copolarized phase difference and the copolarized correlation coefficient. As dual-pol SAR imagery is now available on a regular basis from Cosmo Skymed and TerraSAR-X, and quad-pol data are available from RADARSAT-2, polarimetric SAR can now be utilized on a more regular basis. Optical data from sensors like Aqua MODIS and ENVISAT MERIS can be a useful supplement under certain cloud-free conditions.

• Salberg, Arnt Børre; Rudjord, Øystein & Solberg, Anne H Schistad (2012). Model based oil spill detection using polarimetric SAR. IEEE Geoscience and Remote Sensing Letters. ISSN 1545-598X. p. 5884–5887. doi: 10.1109/IGARSS.2012.6352270. Full text in Research Archive
• Berthelot, Angélique; Solberg, Anne H S; Jarsve, Erlend Morisbak & Gelius, Leiv-J. (2011). Salt diapirs without well defined boundaries - a feasibility study of semi-automatic detection. Geophysical Prospecting. ISSN 0016-8025. 59, p. 682–696. doi: 10.1111/j.1365-2478.2011.00950.x. Full text in Research Archive
• Solberg, Anne H S; Berthelot, Angélique & Gelius, Leiv-J. (2011). Gabor Filters for Segmentation of Salt Structures, Extended Abstracts DVD of the 73rd EAGE Conference & Exhibition incorporating SPE EUROPEC. European Association of Geoscientists and Engineers (EAGE). ISSN 9789073834125. Full text in Research Archive
• Solberg, Anne H S & Gelius, Leiv-J. (2011). New texture attributes from local 2D Fourier spectra. Extended Abstract. In Koster, Klaas (Eds.), Society of Exploration Geophysicists International Exposition and 81st Annual Meeting 2011 (SEG San Antonio 2011). Curran Associates, Inc.. ISSN 9781618391841. p. 1165–1169. doi: 10.1190/1.3627409. Full text in Research Archive
• Berthelot, Angèlique; Solberg, Anne H S & Gelius, Leiv-J. (2010). Attributes designed for detection of salt diapirs without well defined boundaries. In EAGE, 2010 (Eds.), 72nd EAGE Conference & Exhibition incorporating SPE EUROPEC 2010 Barcelona. European Association of Geoscientists and Engineers (EAGE). ISSN 9789073781863. Full text in Research Archive
• Jensen, Are Fritz Charles; Loog, Marco & Solberg, Anne H S (2010). Using Multiscale Spectra in Regularizing Covariance Matrices for Hyperspectral Image Classification. IEEE Transactions on Geoscience and Remote Sensing. ISSN 0196-2892. 48(4), p. 1851–1859. doi: 10.1109/TGRS.2009.2036842. Full text in Research Archive
• Berthelot, Angelique; Gelius, Leiv-J. & Solberg, Anne H S (2009). Use of seismic attributes in auto-tracking of salt structures - A feasibility study. In Mogi, Toru (Eds.), The 9th SEGJ International Symposium Imaging and Interpretation- Science and Technology for Sustainable Development -. Society of Exploration Geophysicists of Japan (SEGJ). ISSN 9784938493066. Full text in Research Archive
• Jensen, Are Fritz Charles; Berge, Asbjørn & Solberg, Anne H S (2008). Regression Approaches to Small Sample Inverse Covariance Matrix Estimation for Hyperspectral Image Classification. IEEE Transactions on Geoscience and Remote Sensing. ISSN 0196-2892. 46(10), p. 2814–2822. doi: 10.1109/TGRS.2008.2001169. Full text in Research Archive
• Brekke, Camilla & Solberg, Anne H S (2008). Classifiers and confidence estimation for oil spill detection in ENVISAT ASAR images. IEEE Geoscience and Remote Sensing Letters. ISSN 1545-598X. 5. doi: 10.1109/LGRS.2007.907174. Full text in Research Archive
• Berge, Asbjørn & Solberg, Anne H S (2007). Improving hyperspectral classifiers: the difference between reducing data dimensionality and reducing classifier parameter complexity. Lecture Notes in Computer Science (LNCS). ISSN 0302-9743. p. 293–302. Full text in Research Archive
• Solberg, Anne H S; Berge, Asbjørn & Jensen, Are Fritz Charles (2007). Robust classification of hyperspectral images. In Bruzzone, Lorenzo (Eds.), Image and Signal Processing for Remote Sensing XIII. SPIE - The International Society for Optics and Photonics. ISSN 9780819469069. Full text in Research Archive
• Solberg, Anne Helene Schistad; Brekke, C & Husøy, Per Ove (2007). Oil spill detection in Radarsat and Envisat SAR images. IEEE Transactions on Geoscience and Remote Sensing. ISSN 0196-2892. 45. doi: 10.1109/TGRS.2006.887019. Full text in Research Archive Show summary

  We present algorithms for automatic detection of oil spills in synthetic aperture radar (SAR) images. The algorithms consist of three main parts, namely: 1) detection of dark spots; 2) feature extraction from the dark spot candidates; and 3) classification of dark spots as oil spills or look-alikes. The algorithms have been trained on a large number of Radarsat and Envisat Advanced Synthetic Aperture Radar (ASAR) images. The performance of the algorithm is compared to manual and semiautomatic approaches in a benchmark study using 59 Radarsat and Envisat images. The algorithms can be considered to be a good alternative to manual inspection when large ocean areas are to be inspected

• Berge, Asbjørn; Jensen, Are Fritz Charles & Solberg, Anne H S (2007). Sparse inverse covariance estimates for hyperspectral image classification. IEEE Transactions on Geoscience and Remote Sensing. ISSN 0196-2892. 45. doi: 10.1109/TGRS.2007.892598. Full text in Research Archive
• Brekke, Camilla; Solberg, Anne H S & Storvik, Geir Olve (2007). Classifying Oil Spills and Look-Alikes in Envisat ASAR Images. In Lacoste, H & Ouwehand, L (Ed.), Proceedings of the Envisat Symposium 2007 - Montreux, Switzerland. European Space Agency. Full text in Research Archive Show summary

  Egentlig ISBN er 92-9291-200-1

• Jensen, Are Fritz Charles & Solberg, Anne H S (2007). Fast Hyperspectral Feature Reduction Using Piecewise Constant Function Approximations. IEEE Geoscience and Remote Sensing Letters. ISSN 1545-598X. 4(4), p. 547–551. doi: 10.1109/LGRS.2007.896331. Full text in Research Archive
• Berge, Asbjørn & Solberg, Anne H S (2006). Structured Gaussian components for hyperspectral image classification. IEEE Transactions on Geoscience and Remote Sensing. ISSN 0196-2892. 44, p. 3386–3396. Full text in Research Archive
• Berge, Asbjørn & Solberg, Anne H S (2006). Sparse covariance estimates for high dimensional classification using the Cholesky decomposition. Lecture Notes in Computer Science (LNCS). ISSN 0302-9743. 4109, p. 835–843. Full text in Research Archive
• Solberg, Anne H S (2006). Data fusion for remote sensing applications. In Chen, Chi (Eds.), Signal and Image Processing for Remote Sensing. Taylor & Francis. ISSN 9780849350917. p. 515–537. Full text in Research Archive
• Solberg, Anne H S & Brekke, Camilla (2006). Classification Methods for Oil Spill Detection in ENVISAT ASAR images. In Bruzzone, Lorenzo (Eds.), Image and Signal Processing for Remote Sensing. SPIE - The International Society for Optics and Photonics. ISSN 9780819464606. Full text in Research Archive
• Brekke, Camilla & Solberg, Anne H S (2005). Oil spill detection by satellite remote sensing. Remote Sensing of Environment. ISSN 0034-4257. 95(1), p. 1–13. Full text in Research Archive
• Storvik, Geir Olve; Fjørtoft, Roger Sverre & Solberg, Anne H S (2005). A Bayesian approach to classification of multi-scale remote sensing data. IEEE Transactions on Geoscience and Remote Sensing. ISSN 0196-2892. 43(3), p. 539–547. doi: 10.1109/TGRS.2004.841395. Full text in Research Archive Show summary

  Several earth observation satellites acquire image bands with different spatial resolutions, e.g., a panchromatic band with high resolution and spectral bands with lower resolution. Likewise, we often face the problem of different resolutions when performing joint analysis of images acquired by different satellites. This work presents models and methods for classification of multiresolution images. The approach is based on the concept of a reference resolution, corresponding to the highest resolution in the dataset. Prior knowledge about the spatial characteristics of the classes is specified through a Markov random field model at the reference resolution. Data at coarser scales are modeled as mixed pixels by relating the observations to the classes at the reference resolution. A Bayesian framework for classification based on this multiscale model is proposed. The classification is realized by an iterative conditional modes (ICM) algorithm. The parameter estimation can be based both on a training set and on pixels with unknown class. A computationally efficient scheme based on a combination of the ICM and the expectation-maximization algorithm is proposed. Results obtained on simulated and real satellite images are presented.

• Aurdal, Lars; Huseby, Ragnar Bang; Eikvil, Line; Solberg, Rune; Schuler, Dagrun Vikhamar & Solberg, Anne H Schistad (2005). Use of hidden Markov models and phenology for multitemporal satellite image classification: applications to mountain vegetation classification. In King, Roger L. & Younan, Nicolas H. (Ed.), Proceedings of the 2005 International Workshop on the Analysis of Multi-Temporal Remote Sensing Images, Biloxi, Mississippi, USA, 16-18 May 2005. IEEE (Institute of Electrical and Electronics Engineers). ISSN 9780780391185. p. 220–224. doi: 10.1109/AMTRSI.2005.1469877. Full text in Research Archive
• Huseby, Ragnar Bang; Aurdal, Lars; Eikvil, Line; Solberg, Rune; Schuler, Dagrun Vikhamar & Solberg, Anne H Schistad (2005). Alignment of growth seasons from satellite data. In King, Roger L. & Younan, Nicolas H. (Ed.), Proceedings of the 2005 International Workshop on the Analysis of Multi-Temporal Remote Sensing Images, Biloxi, Mississippi, USA, 16-18 May 2005. IEEE (Institute of Electrical and Electronics Engineers). ISSN 9780780391185. p. 213–216. doi: 10.1109/AMTRSI.2005.1469875. Full text in Research Archive
• Brekke, C & Solberg, Anne H S (2005). Feature extraction for oil spill detection based on SAR images. Lecture Notes in Computer Science (LNCS). ISSN 0302-9743. 3540. Full text in Research Archive
• Solberg, Anne H S (2004). Flexible nonlinear contextual classification. Pattern Recognition Letters. ISSN 0167-8655. 25(13), p. 1501–1508. Full text in Research Archive

View all works in NVA

• Solberg, Anne H. Schistad (2006). Data fusion for remote sensing applications, in Signal and Image Processing for Remote Sensing (ed. C. H. Chen). Taylor and Francis Books. 24 p. Full text in Research Archive

View all works in NVA

• Singh, Durgesh Kumar; Thomas, Sarina; Cao, Qing; Kampffmeyer, Michael; Jenssen, Robert & Solberg, Anne H Schistad (2026). Semi-Supervised nnU-Net with Weak–Strong Consistency and Boundary Refinement for Mitral Valve Segmentation. doi: https:/papers.miccai.org/miccai-2026-sat/paper/MVAA_089.pdf. Full text in Research Archive Show summary

  Analysis of the mitral valve begins with delineating the valve in the image, a step that is usually performed manually and is therefore slow and observer-dependent. The MICCAI 2026 Mitral Valve Anatomy Analysis (MVAA) challenge defines three segmentation tasks based on cardiac computed tomography (CT), three-dimensional transesophageal echocardiography (3D TEE), and intraoperative surgical video. The unpaired images differ in dimensionality, appearance, and label formulation. Therefore, a separate model is trained for each task within a common self-configuring nnU-Net framework. Labeled images are supervised with soft Dice loss and cross-entropy. Since mitral valve structures are thin, these region losses are complemented by a scheduled boundary loss averaged over the available foreground classes. Expert labels are scarce for cardiac CT and surgical video, while many unlabeled images are available. These unlabeled images are used through FixMatch-style weak–strong consistency learning, where a weak view without intensity augmentation generates confidence-filtered pseudo-labels for intensity-augmented strong views. On the private test set held by the challenge organizers, the proposed approach achieves Dice similarity coefficients of 0.83 for cardiac CT, 0.81 for 3D TEE, and 0.55 for surgical video. Hausdorff distance and average surface distance are reported as boundary measures. The code is publicly available at https://github.com/Durgesh93/MVAA.git.

• Vedal, Amund Hansen; Eikvil, Line; Solberg, Anne H Schistad; Salberg, Arnt Børre & Malde, Ketil (2026). Hierarchical classification and embedding quality of plankton. Full text in Research Archive Show summary

  Poster

• Jenssen, Robert; Eikvil, Line; Solberg, Anne H Schistad; Solheim, Inger & Bjørklund, Petter (2026). Visual Intelligence Annual Report 2025. Full text in Research Archive
• Jenssen, Robert; Eikvil, Line; Solberg, Anne H Schistad; Solheim, Inger & Bjørklund, Petter (2025). Visual Intelligence Annual Report 2024. Full text in Research Archive Show summary

  The Visual Intelligence Annual Report 2024 highlights the centre's progress, activities and achieved innovations for 2022. It describes new deep learning methods which address pressing societal needs in the fields of medicine and health, marine science, the energy sector, and earth observation.

• Jenssen, Robert; Eikvil, Line; Solberg, Anne H Schistad; Solheim, Inger & Bjørklund, Petter (2024). Visual Intelligence Annual Report 2023. Full text in Research Archive
• Aasan, Marius; Kolbjørnsen, Odd; Solberg, Anne H Schistad & Rivera, Adín Ramírez (2024). A Spitting Image: Modular Superpixel Tokenization in Vision Transformers. Full text in Research Archive
• Aasan, Marius; Kolbjørnsen, Odd; Solberg, Anne H Schistad & Rivera, Adín Ramírez (2024). A Spitting Image: Modular Tokenization in Vision Transformers. Full text in Research Archive
• Aasan, Marius; Kolbjørnsen, Odd; Solberg, Anne H Schistad & Rivera, Adín Ramírez (2024). A Spitting Image: Modular Superpixel Tokenization in Vision Transformers. Full text in Research Archive
• Jenssen, Robert; Eikvil, Line; Solberg, Anne H Schistad & Solheim, Inger (2023). Visual Intelligence Annual Report 2022. Full text in Research Archive
• Thomas, Sarina; Andreassen, Børge Solli; Tiago, Cristiana Ferreira; Aase, Svein Arne; Steen, Erik Normann & Šprem, Jurica [Show all 8 contributors for this article] (2023). Graph Convolutional Neural Networks for Automated Echocardiography View Recognition: A Holistic Approach. Full text in Research Archive
• Solberg, Anne H Schistad (2023). Panel discussion - Revolutionizing health care with AI-assisted medical imaging analysis. Full text in Research Archive
• Solberg, Anne H Schistad (2023). Learning from image data. Full text in Research Archive
• Jenssen, Robert; Eikvil, Line; Solberg, Anne H Schistad & Solheim, Inger (2022). Visual Intelligence Annual Report 2021. Full text in Research Archive Show summary

  The Visual Intelligence Annual Report 2021 highlights the centre's progress, activities and achieved innovations for 2021. It describes new deep learning methods which address pressing societal needs in the fields of medicine and health, marine science, the energy sector, and earth observation.

• Aasan, Marius; Rivera, Adín Ramírez; Kolbjørnsen, Odd & Solberg, Anne H Schistad (2022). An Information Metric for Partitions on Images using Lattice Theoretic Entropy. Full text in Research Archive
• Jenssen, Robert; Eikvil, Line; Solberg, Anne H Schistad & Solheim, Inger (2021). Visual Intelligence Annual Report 2020. Full text in Research Archive Show summary

  The Visual Intelligence Annual Report 2020 highlights the centre's progress, activities and achieved innovations for 2020. It describes new deep learning methods which address pressing societal needs in the fields of medicine and health, marine science, the energy sector, and earth observation.

• Jenssen, Robert; Solberg, Anne H Schistad & Eikvil, Line (2021). Norge er verdensledende på bildeanalyse med kunstig intelligens. Full text in Research Archive
• Di, Haibin; Zhao, Tao; Jayaram, Vikram; Wu, Xinming; Huang, Lei & AlRegib, Ghassan [Show all 15 contributors for this article] (2019). Introduction to special section: Machine learning in seismic data analysis. Interpretation. ISSN 2324-8858. 7(3), p. SEI–SEII. doi: 10.1190/INT-2019-0609-SPSEINTRO.1. Full text in Research Archive
• Waldeland, Anders Ueland & Solberg, Anne H Schistad (2017). Salt Classification Using Deep Learning. Full text in Research Archive
• Waldeland, Anders Ueland & Solberg, Anne H Schistad (2016). 3D Attributes and Classification of Salt Bodies on Unlabelled Datasets. Full text in Research Archive
• Hverven, Stine Myhre; Austeng, Andreas & Solberg, Anne H Schistad (2016). Adaptive Beamforming on Muscle Fiber Features. Full text in Research Archive
• Solberg, Anne H Schistad; Gelius, Leiv-J.; Jensen, Are Charles & Heeremans, Michel (2012). Phase attributes by considering local 3D structure orientation. Full text in Research Archive
• Berthelot, Angélique; Solberg, Anne H Schistad; Morisbak, Erlend & Gelius, Leiv-J. (2012). 3D segmentation of salt using texture attributes. Full text in Research Archive
• Salberg, Arnt Børre; Rudjord, Øystein & Solberg, Anne H S (2012). Oil spill detection in compact polarimetry SAR images. Full text in Research Archive
• Jensen, Are Fritz Charles; Berge, Asbjørn & Solberg, Anne H S (2007). Regression Approaches to Small Sample Inverse Covariance Matrix Estimation for Hyperspectral Image Classification. Full text in Research Archive
• Solberg, Anne H S & Brekke, Camilla (2006). Automatisk varsling av oljesøl. [Journal]. Apollon. Full text in Research Archive
• Solberg, Anne H S (2006). Effektiv varsling av oljesøl. [Journal]. Havneavisen. Full text in Research Archive
• Berge, Asbjørn & Solberg, Anne H S (2006). Sparse covariance estimates for hyperspectral image classification using the Cholesky decomposition. Full text in Research Archive
• Brekke, Camilla & Solberg, Anne H S (2006). Classification Methods for Oil Spill Detection in ENVISAT ASAR Images. Full text in Research Archive
• Brekke, Camilla & Solberg, Anne H. Schistad (2006). Classification methods for oil spill detection by SAR imaging. Full text in Research Archive
• Brekke, Camilla & Solberg, Anne H S (2005). Feature extraction for oil spill detection based on SAR images. Full text in Research Archive
• Solberg, Anne H S (2005). Automatic oil spill detection based on ENVISAT, RADARSAT and ERS images. Full text in Research Archive
• Jensen, Tine; Solberg, Anne & Ulvik, Oddbjørg Skjær (2005). Doing research with children in vulnerable situations: Methodological and ethical challenges. Full text in Research Archive
• Solberg, Anne H S (2004). Oljesøldeteksjon. Full text in Research Archive
• Andreassen, Børge Solli; Samset, Eigil; Solberg, Anne H Schistad; Thomas, Sarina & Volgyes, David (2025). Deep Learning for Mitral Annulus Delineation in 3D Transesophageal Echocardiography. Universitetet i Oslo. Full text in Research Archive Show summary

  Valvular heart disease is a significant health concern, particularly in aging populations, with mitral valve disease being among the most common. Three-dimensional transesophageal echocardiography (TEE) provides high-quality imaging for assessing the mitral valve. Delineating the mitral annulus in 3D images is an important step for accurate quantitative morphological assessment of the valve. However, manual analysis is time-consuming and prone to interobserver variability. Deep learning is a class of machine learning methods that has led to significant progress across an extensive range of applications. In medical imaging, it has enabled improvements across a wide array of tasks, including classification, segmentation, and landmark detection. In this thesis, deep learning is applied to automate the delineation of the mitral annulus from 3D TEE images. The thesis presents a progression of methods across three papers, each building on the last, gradually incorporating more anatomical context and spatial structure. The first paper applies a 2D model with post-processing to enforce annular continuity. The second paper introduces a periodic 3D model using cylindrical coordinates, enabling direct segmentation and prediction of anatomical orientation. The third paper integrates anatomical constraints through a graph-based model to improve the robustness and consistency of the predicted annulus. The methods were evaluated on annotated 3D TEE datasets, showing accurate and consistent delineation of the mitral annulus across multiple subjects. In addition to the specific application, the thesis proposes a novel loss function and coordinate decoding strategy for deep learning with potential applicability in landmark detection tasks, both in medical and natural images. These Developments support the potential for integration into practical workflows, which could assist clinicians in assessing mitral valves while contributing to automation in cardiac imaging.

• Rindal, Ole Marius Hoel; Austeng, Andreas & Solberg, Anne H Schistad (2019). Software Beamforming in Medical Ultrasound Imaging - a blessing and a curse. Universitetet i Oslo. Full text in Research Archive Show summary

  Medical ultrasound (US) imaging is a non-invasive imaging modality. Smaller and cheaper US systems make US imaging available to more people, leading to a democratization of medical US imaging. The improvements of general processing hardware allow the reconstruction of US images to be done in software. These implementations are known as software beamforming and provide access to the US data earlier in the processing chain. Adaptive beamforming exploits the early access to the full US data with algorithms adapting the processing to the data. Adaptive beamforming claims improved image quality. The improved image will potentially result in an improved diagnosis. Adaptive beamformers have seen enormous popularity in the research community with exponential growth in the number of papers published. However, the complexity of the algorithms makes them hard to re-implement, making a thorough comparison of the algorithms difficult. The UltraSound ToolBox (USTB https://www.USTB.no) is an open source processing framework facilitating the comparison of imaging techniques and the dissemination of research results. The USTB, including the implementation of several state-of-the-art adaptive beamformers, has partly been developed in this thesis and used to produce most of the results presented. The results show that some of the contrast improvements reported in the literature turn out to be from secondary effects of adaptive processing. More specifically, we show that many state-of-the-art algorithms alter the dynamic range. These dynamic range alterations are invalidating the conventional contrast metrics. Said differently; many adaptive algorithms are so flexible that they instead of improving the image quality are merely optimizing the metrics used to evaluate the image quality. We suggest a dynamic range test, compromising data, and code, to assess whether an algorithm alters the dynamic range. A thorough review of the contrast metrics used in US imaging shows there is no consensus on the metrics used in the research literature. Therefore, our introduction of the generalized contrast to noise ratio (GCNR) is essential since this is a contrast metric immune to dynamic range alterations. The GCNR is a remedy for the curse of the metric breaking abilities of software beamforming. Software beamforming also has its blessings. The flexible implementations made possible by software beamforming does lead to improved image quality. The improved resolution of the minimum variance adaptive beamformer does lead to enhanced visualization of the interventricular septum in the human heart. The ability to do beamforming in software allows the implementation of the full reconstruction chain from raw data to the final rendered images on an iPhone. As well as the results presented in the published papers, this thesis does a thorough review of the software beamforming processing chain as implemented in the USTB.

• Waldeland, Anders U.; Solberg, Anne H Schistad & Gelius, Leiv-J. (2018). Seismic image analysis for applications related to iterative 3D velocity model building. Universitetet i Oslo. Full text in Research Archive
• Salberg, Arnt Børre; Longépé, Nicolas; Hansen, Morten Wergeland; Rudjord, Øystein; Larsen, Siri Øyen & Solberg, Anne H Schistad [Show all 8 contributors for this article] (2013). Report SAR analysis and comparisons - EU FP-7 project SeaU, Deliverable 12.1. Research Executive Agency - European Commission. Full text in Research Archive
• Solberg, Anne H S & Trier, Øivind Thorvald Due (2011). Radarsat-2 oil spill detection. Performance study 2010. Norsk Regnesentral. Full text in Research Archive
• Salberg, Arnt-Børre; Solberg, Svein; Solberg, Anne H. Schistad & Lange, Holger (2009). Leaf Area Index Estimation from Land Cover Classification Based on Hyperspectral Data. Norsk Regnesentral. Full text in Research Archive
• Solberg, Anne H. Schistad & Trier, Øivind Thorvald Due (2009). SatHavOlje. Improvements during the 2008 project. Norsk Regnesentral. Full text in Research Archive
• Solberg, Anne H. Schistad (2008). Confidence Estimation - Report for SatHav Olje. Norsk Regnesentral. Full text in Research Archive
• Huseby, Ragnar Bang & Solberg, Anne H. Schistad (2008). Statistical analysis of oil spill data. Norsk Regnesentral. Full text in Research Archive

View all works in NVA