AMD Severity Prediction And Explainability Using Image Registration And Deep Embedded Clustering

Dwarikanath Mahapatra

arxiv: 1907.03075 · v1 · pith:36GR7LJ6new · submitted 2019-07-06 · 💻 cs.CV · eess.IV

AMD Severity Prediction And Explainability Using Image Registration And Deep Embedded Clustering

Dwarikanath Mahapatra This is my paper

Pith reviewed 2026-05-25 01:44 UTC · model grok-4.3

classification 💻 cs.CV eess.IV

keywords AMDOCTimage registrationdeep embedded clusteringseverity predictionexplainabilitydeep learningretinal imaging

0 comments

The pith

Deep learning registration and clustering predicts AMD severity from OCT images without a standard clinical scale.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper proposes using image registration and deep embedded clustering to predict the severity of age-related macular degeneration from optical coherence tomography scans. It shows that this approach achieves classification performance comparable to existing state-of-the-art methods while also performing well on new data. Registration outputs are claimed to provide clearer explanations for the predictions than standard class activation maps. This matters because AMD lacks a uniform severity scale, so the method offers a data-driven way to assess and explain disease progression.

Core claim

By combining deep learning based image registration with deep embedded clustering, the method identifies diseased cases and predicts their severity levels in AMD OCT images, matching state-of-the-art classification accuracy and offering improved explainability through registration outputs compared to class activation maps.

What carries the argument

Deep embedded clustering applied to registered OCT images to group cases by inferred severity levels.

If this is right

Classification performance matches state of the art methods.
Predicted severity performs well on previously unseen data.
Registration output provides better explainability than class activation maps for label and severity decisions.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

Such methods could enable consistent severity tracking across different clinics despite varying clinical practices.
Future work might test if these clusters align with actual patient outcomes like vision loss rates.
Integration with longitudinal data could allow prediction of disease progression rates.

Load-bearing premise

That the clusters identified by deep embedded clustering correspond to meaningful differences in AMD severity even without a standard clinical scale.

What would settle it

A comparison showing that the predicted severity levels do not align with expert clinical assessments or patient outcomes on an independent test set.

Figures

Figures reproduced from arXiv: 1907.03075 by Dwarikanath Mahapatra.

**Figure 1.** Figure 1: Architecture of our proposed network for AMD classification and severity estimation. A regression network for image registration and deep embedded clustering network are combined to achieve our objectives. Thus the following mapping between image labels and cluster labels are obtained Normal ∈ {1, 2, 3}, DME ∈ {4, 5, 6, 7}, and AMD ∈ {8, 9, 10}. 2.5 Predicting Severity of test image When a test image come… view at source ↗

**Figure 2.** Figure 2: third column shows examples of normal images that were rightly classified by Reg − DEC but incorrectly classified as AMD by DenseNet. The green regions highlight disease activty as identified by DenseNet, which is erroneous since there are no abnormalities here. Reg − DEC does not show any localization of pathologies in these examples. The fourth column shows examples of DME that were rightly identified b… view at source ↗

read the original abstract

We propose a method to predict severity of age related macular degeneration (AMD) from input optical coherence tomography (OCT) images. Although there is no standard clinical severity scale for AMD, we leverage deep learning (DL) based image registration and clustering methods to identify diseased cases and predict their severity. Experiments demonstrate our approach's disease classification performance matches state of the art methods. The predicted disease severity performs well on previously unseen data. Registration output provides better explainability than class activation maps regarding label and severity decisions

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

This is an application of standard registration and clustering to AMD on OCT with no new methods and no external validation for the derived severity labels.

read the letter

The main takeaway is that this is an application paper using existing deep registration and clustering techniques on OCT images for AMD severity, without new methods, and the severity labels come from the clustering without external clinical validation to back them up. The paper does a reasonable job of addressing explainability by suggesting that registration outputs are better than class activation maps for understanding the decisions on labels and severity. It also tries to work around the absence of a standard severity scale by leveraging unsupervised clustering, which is a practical move for this domain. Where it falls short is in the grounding of the results. The strongest claims about matching state of the art and performing well on unseen data can't be assessed because the abstract provides no dataset details, no quantitative metrics, and no description of how the clusters are tied to actual severity. The stress-test note is accurate: without an independent anchor like doctor-assigned grades or other clinical metrics, the internal consistency of the clusters doesn't translate to reliable severity prediction. This makes the generalization claim hard to trust. The approach follows standard practices in the literature, so the citation pattern is fine and there's no sign of circular fitting presented as something else. But the evidence is too thin to support the conclusions as stated. This paper is aimed at researchers in medical computer vision working on retinal diseases. Someone interested in quick applications of DL tools to eye imaging might find it useful as a case study, but it won't offer much to those seeking methodological advances or clinically validated tools. I wouldn't bring it to the next reading group, and I wouldn't cite it in my own work. It does not deserve peer review in this form because the core issue with validating the severity predictions against real clinical measures needs to be addressed first.

Referee Report

2 major / 1 minor

Summary. The paper proposes using deep learning-based image registration combined with deep embedded clustering on OCT images to identify AMD cases and predict their severity levels, in the absence of a standard clinical scale. It claims that the disease classification performance matches state-of-the-art methods, that the severity predictions generalize well to unseen data, and that registration outputs provide superior explainability over class activation maps for both labels and severity decisions.

Significance. If the clusters derived from deep embedded clustering can be shown to align with clinically meaningful AMD severity distinctions and the registration-based explanations hold under external validation, the approach could enable quantitative severity assessment and improved interpretability in a domain lacking standardized scales. The reported matching of SOTA classification performance and generalization would be a useful contribution if substantiated with full experimental details.

major comments (2)

[Abstract] Abstract: the central claim that 'the predicted disease severity performs well on previously unseen data' and matches SOTA classification performance is unsupported because the abstract (and available description) supplies no dataset details, metrics, baselines, validation procedures, or cross-validation scheme, preventing any assessment of whether the data actually supports the claims.
[Abstract] Abstract/Methods (clustering pipeline): the assertion that clusters meaningfully encode AMD severity levels lacks an external clinical validation anchor (e.g., correlation to expert grades, visual acuity, or lesion counts). Since the paper notes there is no standard clinical severity scale, deriving severity labels directly from the fitted clusters creates a circularity risk that undermines the generalization and explainability claims.

minor comments (1)

[Abstract] The abstract refers to 'registration output' for explainability but does not specify the registration algorithm, loss terms, or how alignment outputs are mapped to severity decisions versus CAMs.

Simulated Author's Rebuttal

2 responses · 0 unresolved

Thank you for reviewing our manuscript. We address each of the major comments below.

read point-by-point responses

Referee: [Abstract] Abstract: the central claim that 'the predicted disease severity performs well on previously unseen data' and matches SOTA classification performance is unsupported because the abstract (and available description) supplies no dataset details, metrics, baselines, validation procedures, or cross-validation scheme, preventing any assessment of whether the data actually supports the claims.

Authors: The abstract provides a high-level overview and is constrained by word limits. The full manuscript details the OCT dataset used, the performance metrics (such as accuracy and AUC), the state-of-the-art baselines compared against, and the validation procedures including the cross-validation scheme in the Experiments section. These support the claims made. We will revise the abstract to include brief mention of key results and dataset size to address this concern. revision: partial
Referee: [Abstract] Abstract/Methods (clustering pipeline): the assertion that clusters meaningfully encode AMD severity levels lacks an external clinical validation anchor (e.g., correlation to expert grades, visual acuity, or lesion counts). Since the paper notes there is no standard clinical severity scale, deriving severity labels directly from the fitted clusters creates a circularity risk that undermines the generalization and explainability claims.

Authors: We note that the lack of a standard clinical severity scale for AMD is stated in the paper and is the reason for adopting an unsupervised clustering approach to derive severity levels from the data. The clusters are obtained via deep embedded clustering on registered images, and we show that the severity predictions generalize to unseen data while the registration provides visual explanations. This avoids circularity because no pre-existing severity labels are used; the method discovers structure in the data. External validation with clinical anchors is a valuable direction for future work but is outside the scope of the current technical contribution focused on the registration and clustering pipeline. revision: no

Circularity Check

1 steps flagged

Severity labels and predictions both derive from the same deep embedded clustering step

specific steps

self definitional [Abstract]
"Although there is no standard clinical severity scale for AMD, we leverage deep learning (DL) based image registration and clustering methods to identify diseased cases and predict their severity. ... The predicted disease severity performs well on previously unseen data."

Severity is introduced solely via the clustering output (no external scale exists), yet the same clustering procedure is then presented as producing independent 'predictions' whose quality is evaluated on held-out data. The performance metric therefore measures how well the clustering reproduces its own assignments rather than matching any external clinical quantity.

full rationale

The paper explicitly notes the absence of any external clinical severity scale and instead uses deep embedded clustering both to define severity categories and to generate the 'predictions' on unseen data. This makes the reported performance an internal consistency check on the clustering itself rather than an independent validation against clinical ground truth. No equations or self-citations are needed to see the reduction; the abstract states the construction directly.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review provides no details on free parameters, axioms, or invented entities; ledger left empty.

pith-pipeline@v0.9.0 · 5600 in / 1040 out tokens · 23260 ms · 2026-05-25T01:44:21.695301+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

75 extracted references · 75 canonical work pages · 3 internal anchors

[1]

https://www.eyepacs.com

work page
[2]

”the insight segmentation and registration toolkit” www.itk.org

work page
[3]

In: Proc

Balakrishnan, G., Zhao, A., Sabuncu, M., Guttag, J.: An supervised learning model for deformable medical image registration. In: Proc. CVPR. pp. 9252–9260 (2018)

work page 2018
[4]

Computer Vision and Image Understanding 184, 57–65 (2019)

Bozorgtabar, B., Mahapatra, D., von Teng, H., Pollinger, A., Ebner, L., Thiran, J.P., M.Reyes.: Informative sample generation using class aware generative ad- versarial networks for classiﬁcation of chest xrays. Computer Vision and Image Understanding 184, 57–65 (2019)

work page 2019
[5]

Mahapatra, B.B., Garnavi, R.: Image super-resolution using progressive gen- erative adversarial networks for medical image analysis

D. Mahapatra, B.B., Garnavi, R.: Image super-resolution using progressive gen- erative adversarial networks for medical image analysis. Computerized Medical Imaging and Graphics 71(1), 30–39 (2019)

work page 2019
[6]

al.: Flownet: Learning optical ﬂow with convolu- tional networks

Dosovitskiy, A., Fischer, P., et. al.: Flownet: Learning optical ﬂow with convolu- tional networks. In: In Proc. IEEE ICCV. pp. 2758–2766 (2015)

work page 2015
[7]

In: In Proc

Graziani, M., Andrearczyk, V., Mller, H.: Regression concept vectors for bidirec- tional explanations in histopathology. In: In Proc. MICCAI-iMIMIC. pp. 124–132 (2018)

work page 2018
[8]

In: In Proc

He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. In: In Proc. CVPR (2016)

work page 2016
[9]

Densely Connected Convolutional Networks

Huang, G., Liu, Z., van der Maaten, L., Weinberger, K.: Densely connected con- volutional networks. In: https://arxiv.org/abs/1608.06993, (2016)

work page internal anchor Pith review Pith/arXiv arXiv 2016
[10]

In: In Proc: ACM Intl

Kuang, H., Guthier, B., Saini, M., Mahapatra, D., Saddik, A.E.: A real-time smart assistant for video surveillance through handheld devices. In: In Proc: ACM Intl. Conf. Multimedia. pp. 917–920 (2014)

work page 2014
[11]

IEEE Trans

Li, Z., Mahapatra, D., J.Tielbeek, Stoker, J., van Vliet, L., Vos, F.: Image registra- tion based on autocorrelation of local structure. IEEE Trans. Med. Imaging 35(1), 63–75 (2016)

work page 2016
[12]

In: In Proc: MICCAI workshop on Image Analysis of Human Brain Devel- opment (IAHBD) (2011)

Mahapatra, D.: Neonatal brain mri skull stripping using graph cuts and shape priors. In: In Proc: MICCAI workshop on Image Analysis of Human Brain Devel- opment (IAHBD) (2011)

work page 2011
[13]

In: Proc

Mahapatra, D.: Cardiac lv and rv segmentation using mutual context information. In: Proc. MICCAI-MLMI. pp. 201–209 (2012)

work page 2012
[14]

In: In Proc: SPIE Medical Imaging (2012) 9

Mahapatra, D.: Groupwise registration of dynamic cardiac perfusion images using temporal information and segmentation information. In: In Proc: SPIE Medical Imaging (2012) 9

work page 2012
[15]

In: Proc

Mahapatra, D.: Landmark detection in cardiac mri using learned local image statis- tics. In: Proc. MICCAI-Statistical Atlases and Computational Models of the Heart. Imaging and Modelling Challenges (STACOM). pp. 115–124 (2012)

work page 2012
[16]

Mahapatra, D.: Skull stripping of neonatal brain mri: Using prior shape information with graphcuts. J. Digit. Imaging 25(6), 802–814 (2012)

work page 2012
[17]

Mahapatra, D.: Cardiac image segmentation from cine cardiac mri using graph cuts and shape priors. J. Digit. Imaging 26(4), 721–730 (2013)

work page 2013
[18]

Mahapatra, D.: Cardiac mri segmentation using mutual context information from left and right ventricle. J. Digit. Imaging 26(5), 898–908 (2013)

work page 2013
[19]

In: Proc

Mahapatra, D.: Graph cut based automatic prostate segmentation using learned semantic information. In: Proc. IEEE ISBI. pp. 1304–1307 (2013)

work page 2013
[20]

Mahapatra, D.: Joint segmentation and groupwise registration of cardiac perfusion images using temporal information. J. Digit. Imaging 26(2), 173–182 (2013)

work page 2013
[21]

Mahapatra, D.: Automatic cardiac segmentation using semantic information from random forests. J. Digit. Imaging. 27(6), 794–804 (2014)

work page 2014
[22]

Computer Vision and Image Understanding 151(1), 114–123 (2016)

Mahapatra, D.: Combining multiple expert annotations using semi-supervised learning and graph cuts for medical image segmentation. Computer Vision and Image Understanding 151(1), 114–123 (2016)

work page 2016
[23]

Pattern Recognition 63(1), 700–709 (2017)

Mahapatra, D.: Semi-supervised learning and graph cuts for consensus based med- ical image segmentation. Pattern Recognition 63(1), 700–709 (2017)

work page 2017
[24]

In: In Proc

Mahapatra, D., Bozorgtabar, S., Hewavitahranage, S., Garnavi, R.: Image super resolution using generative adversarial networks and local saliencymaps for retinal image analysis,. In: In Proc. MICCAI. pp. 382–390 (2017)

work page 2017
[25]

In: In Proc

Mahapatra, D., Bozorgtabar, S., Thiran, J.P., Reyes, M.: Eﬃcient active learning for image classiﬁcation and segmentation using a sample selection and conditional generative adversarial network. In: In Proc. MICCAI (2). pp. 580–588 (2018)

work page 2018
[26]

In: In Proc

Mahapatra, D., Buhmann, J.: Obtaining consensus annotations for retinal image segmentation using random forest and graph cuts. In: In Proc. OMIA. pp. 41–48 (2015)

work page 2015
[27]

In: In Proc

Mahapatra, D., Buhmann, J.: Visual saliency based active learning for prostate mri segmentation. In: In Proc. MLMI. pp. 9–16 (2015)

work page 2015
[28]

SPIE Journal of Medical Imaging 3(1) (2016)

Mahapatra, D., Buhmann, J.: Visual saliency based active learning for prostate mri segmentation. SPIE Journal of Medical Imaging 3(1) (2016)

work page 2016
[29]

In: Proc

Mahapatra, D., Buhmann, J.: Automatic cardiac rv segmentation using semantic information with graph cuts. In: Proc. IEEE ISBI. pp. 1094–1097 (2013)

work page 2013
[30]

IEEE Trans

Mahapatra, D., Buhmann, J.: Analyzing training information from random forests for improved image segmentation. IEEE Trans. Imag. Proc. 23(4), 1504–1512 (2014)

work page 2014
[31]

IEEE Trans

Mahapatra, D., Buhmann, J.: Prostate mri segmentation using learned semantic knowledge and graph cuts. IEEE Trans. Biomed. Engg. 61(3), 756–764 (2014)

work page 2014
[32]

In: In Proc

Mahapatra, D., Buhmann, J.: A ﬁeld of experts model for optic cup and disc segmentation from retinal fundus images. In: In Proc. IEEE ISBI. pp. 218–221 (2015)

work page 2015
[33]

In: In Proc

Mahapatra, D., Ge, Z.: Training data independent image registration with gans using transfer learning and segmentation information. In: In Proc. IEEE ISBI (2019)

work page 2019
[34]

In: In Proc

Mahapatra, D., Ge, Z., Sedai, S., Chakravorty., R.: Joint registration and segmen- tation of xray images using generative adversarial networks. In: In Proc. MICCAI- MLMI. pp. 73–80 (2018) 10

work page 2018
[35]

IEEE Journal of Selected Topics in Signal Processing

Mahapatra, D., Gilani, S., Saini., M.: Coherency based spatio-temporal saliency detection for video object segmentation. IEEE Journal of Selected Topics in Signal Processing. 8(3), 454–462 (2014)

work page 2014
[36]

IEEE Trans

Mahapatra, D., J.Tielbeek, Makanyanga, J., Stoker, J., Taylor, S., Vos, F., Buh- mann, J.: Automatic detection and segmentation of crohn’s disease tissues from abdominal mri. IEEE Trans. Med. Imaging 32(12), 1232–1248 (2013)

work page 2013
[37]

In: Proc

Mahapatra, D., J.Tielbeek, Makanyanga, J., Stoker, J., Taylor, S., Vos, F., Buh- mann, J.: Active learning based segmentation of crohn’s disease using principles of visual saliency. In: Proc. IEEE ISBI. pp. 226–229 (2014)

work page 2014
[38]

In: In Proc: MICCAI-ABD (2014)

Mahapatra, D., J.Tielbeek, Makanyanga, J., Stoker, J., Taylor, S., Vos, F., Buh- mann, J.: Combiningmultiple expert annotations using semi-supervised learning and graph cuts for crohns disease segmentation. In: In Proc: MICCAI-ABD (2014)

work page 2014
[39]

Mahapatra, D., J.Tielbeek, Vos, F., Buhmann, J.: A supervised learning approach for crohn’s disease detection using higher order image statistics and a novel shape asymmetry measure. J. Digit. Imaging 26(5), 920–931 (2013)

work page 2013
[40]

In: In Proc

Mahapatra, D., Li, Z., Vos, F., Buhmann, J.: Joint segmentation and groupwise registration of cardiac dce mri using sparse data representations. In: In Proc. IEEE ISBI. pp. 1312–1315 (2015)

work page 2015
[41]

In: IEEE International Conference on Industrial Technology (ICIT)

Mahapatra, D., Routray, A., Mishra, C.: An active snake model for classiﬁcation of extreme emotions. In: IEEE International Conference on Industrial Technology (ICIT). pp. 2195–2199 (2006)

work page 2006
[42]

In: In Proc

Mahapatra, D., Roy, P., Sedai, S., Garnavi, R.: A cnn based neurobiology inspired approach for retinal image quality assessment. In: In Proc. EMBC. pp. 1304–1307 (2016)

work page 2016
[43]

In: In Proc

Mahapatra, D., Roy, P., Sedai, S., Garnavi, R.: Retinal image quality classiﬁcation using saliency maps and cnns. In: In Proc. MICCAI-MLMI. pp. 172–179 (2016)

work page 2016
[44]

In: In 13th International Conference on Biomedical Engineering (2008)

Mahapatra, D., Roy, S., Sun, Y.: Retrieval of mr kidney images by incorporating spatial information in histogram of low level features. In: In 13th International Conference on Biomedical Engineering (2008)

work page 2008
[45]

In: IEEE ICME

Mahapatra, D., Saini, M., Sun, Y.: Illumination invariant tracking in oﬃce envi- ronments using neurobiology-saliency based particle ﬁlter. In: IEEE ICME. pp. 953–956 (2008)

work page 2008
[46]

In: Proc

Mahapatra, D., Sch¨uﬄer, P., Tielbeek, J., Vos, F., Buhmann, J.: Semi-supervised and active learning for automatic segmentation of crohn’s disease. In: Proc. MIC- CAI, Part 2. pp. 214–221 (2013)

work page 2013
[47]

In: Proc

Mahapatra, D., Sun, Y.: Nonrigid registration of dynamic renal MR images using a saliency based MRF model. In: Proc. MICCAI. pp. 771–779 (2008)

work page 2008
[48]

In: Proc

Mahapatra, D., Sun, Y.: Registration of dynamic renal MR images using neurobi- ological model of saliency. In: Proc. ISBI. pp. 1119–1122 (2008)

work page 2008
[49]

In: In 13th International Conference on Biomedical Engi- neering (2008)

Mahapatra, D., Sun, Y.: Using saliency features for graphcut segmentation of per- fusion kidney images. In: In 13th International Conference on Biomedical Engi- neering (2008)

work page 2008
[50]

In: Proc

Mahapatra, D., Sun, Y.: Joint registration and segmentation of dynamic cardiac perfusion images using mrfs. In: Proc. MICCAI. pp. 493–501 (2010)

work page 2010
[51]

In: Proc

Mahapatra, D., Sun., Y.: An mrf framework for joint registration and segmentation of natural and perfusion images. In: Proc. IEEE ICIP. pp. 1709–1712 (2010)

work page 2010
[52]

In: Proc

Mahapatra, D., Sun, Y.: Retrieval of perfusion images using cosegmentation and shape context information. In: Proc. APSIPA Annual Summit and Conference (ASC) (2010) 11

work page 2010
[53]

EURASIP Journal on Image and Video Processing

Mahapatra, D., Sun, Y.: Rigid registration of renal perfusion images using a neu- robiology based visual saliency model. EURASIP Journal on Image and Video Processing. pp. 1–16 (2010)

work page 2010
[54]

In: Proc

Mahapatra, D., Sun, Y.: A saliency based mrf method for the joint registration and segmentation of dynamic renal mr images. In: Proc. ICDIP (2010)

work page 2010
[55]

IEEE Trans

Mahapatra, D., Sun, Y.: Mrf based intensity invariant elastic registration of cardiac perfusion images using saliency information. IEEE Trans. Biomed. Engg. 58(4), 991–1000 (2011)

work page 2011
[56]

In: In Proc: MICCAI

Mahapatra, D., Sun, Y.: Orientation histograms as shape priors for left ventricle segmentation using graph cuts. In: In Proc: MICCAI. pp. 420–427 (2011)

work page 2011
[57]

IEEE Trans

Mahapatra, D., Sun, Y.: Integrating segmentation information for improved mrf- based elastic image registration. IEEE Trans. Imag. Proc. 21(1), 170–183 (2012)

work page 2012
[58]

In: Proc

Mahapatra, D., Tielbeek, J., Buhmann, J., Vos, F.: A supervised learning based ap- proach to detect crohn’s disease in abdominal mr volumes. In: Proc. MICCAI work- shop Computational and Clinical Applications in Abdominal Imaging(MICCAI- ABD). pp. 97–106 (2012)

work page 2012
[59]

In: Proc

Mahapatra, D., Tielbeek, J., Vos, F., ., J.B.: Crohn’s disease tissue segmentation from abdominal mri using semantic information and graph cuts. In: Proc. IEEE ISBI. pp. 358–361 (2013)

work page 2013
[60]

In: Proc

Mahapatra, D., Tielbeek, J., Vos, F., Buhmann, J.: Localizing and segmenting crohn’s disease aﬀected regions in abdominal mri using novel context features. In: Proc. SPIE Medical Imaging (2013)

work page 2013
[61]

In: Proc

Mahapatra, D., Tielbeek, J., Vos, F., Buhmann, J.: Weakly supervised semantic segmentation of crohn’s disease tissues from abdominal mri. In: Proc. IEEE ISBI. pp. 832–835 (2013)

work page 2013
[62]

In: In Proc

Mahapatra, D., Vos, F., Buhmann, J.: Crohn’s disease segmentation from mri using learned image priors. In: In Proc. IEEE ISBI. pp. 625–628 (2015)

work page 2015
[63]

Computer Methods and Programs in Biomedicine 128(1), 75–85 (2016)

Mahapatra, D., Vos, F., Buhmann, J.: Active learning based segmentation of crohns disease from abdominal mri. Computer Methods and Programs in Biomedicine 128(1), 75–85 (2016)

work page 2016
[64]

In: SPIE HVEI

Mahapatra, D., Winkler, S., Yen, S.: Motion saliency outweighs other low-level features while watching videos. In: SPIE HVEI. pp. 1–10 (2008)

work page 2008
[65]

In: In Proc

Pereira, S., Meier, R., Alves, V., Reyes, M., Silva., C.: Automatic brain tumor grading from mri data using convolutional neural networks and quality assessment. In: In Proc. MICCAI-iMIMIC. pp. 106–114 (2018)

work page 2018
[66]

IEEE Trans

Rasti, R., Rabbani, H., Mehri, A., Hajizadeh, F.: Macular oct classiﬁcation using a multi-scale convolutional neural network ensemble. IEEE Trans. Med. Imag. 37(4), 1024–1034 (2018)

work page 2018
[67]

Very Deep Convolutional Networks for Large-Scale Image Recognition

Simonyan, K., Zisserman., A.: Very deep convolutional networks for large-scale image recognition. CoRR abs/1409.1556 (2014)

work page internal anchor Pith review Pith/arXiv arXiv 2014
[68]

In: MIC- CAI

Sokooti, H., de Vos, B., Berendsen, F., Lelieveldt, B., Isgum, I., Staring, M.: Non- rigid image registration using multiscale 3d convolutional neural networks. In: MIC- CAI. pp. 232–239 (2017)

work page 2017
[69]

Journal of Mach

Vincent, P., Larochelle, H., Lajoie, I., Bengio, Y., Manzagol, P.: Stacked denois- ing autoencoders: Learning useful representations in a deep network with a local denoising criterion. Journal of Mach. Learn. Res. 11, 3371–3408 (2010)

work page 2010
[70]

End-to-End Unsupervised Deformable Image Registration with a Convolutional Neural Network

de Vos, B., Berendsen, F., Viergever, M., Staring, M., Isgum, I.: End-to-end un- supervised deformable image registration with a convolutional neural network. In: arXiv preprint arXiv:1704.06065 (2017) 12

work page internal anchor Pith review Pith/arXiv arXiv 2017
[71]

IEEE EMBC

Vos, F.M., Tielbeek, J., Naziroglu, R., Li, Z., Sch¨uﬄer, P., Mahapatra, D., Wiebel, A., Lavini, C., Buhmann, J., Hege, H., Stoker, J., van Vliet, L.: Computational modeling for assessment of IBD: to be or not to be? In: Proc. IEEE EMBC. pp. 3974–3977 (2012)

work page 2012
[72]

In: Proc

Xie, J., Girshick, R., Farhadi, A.: Unsupervised deep embedding for clustering analysis. In: Proc. ICML. pp. 478–487 (2016)

work page 2016
[73]

In: Proc

Zhou, B., Khosla, A., Lapedriza, A., Oliva, A., Torralba, A.: Learning deep features for discriminative localization. In: Proc. CVPR. pp. 2921–2929 (2016)

work page 2016
[74]

In: In Proc

Zilly, J., Buhmann, J., Mahapatra, D.: Boosting convolutional ﬁlters with entropy sampling for optic cup and disc image segmentation from fundus images. In: In Proc. MLMI. pp. 136–143 (2015)

work page 2015
[75]

In Press Computerized Medical Imaging and Graphics 55(1), 28–41 (2017)

Zilly, J., Buhmann, J., Mahapatra, D.: Glaucoma detection using entropy sampling and ensemble learning for automatic optic cup and disc segmentation. In Press Computerized Medical Imaging and Graphics 55(1), 28–41 (2017)

work page 2017

[1] [1]

https://www.eyepacs.com

work page

[2] [2]

”the insight segmentation and registration toolkit” www.itk.org

work page

[3] [3]

In: Proc

Balakrishnan, G., Zhao, A., Sabuncu, M., Guttag, J.: An supervised learning model for deformable medical image registration. In: Proc. CVPR. pp. 9252–9260 (2018)

work page 2018

[4] [4]

Computer Vision and Image Understanding 184, 57–65 (2019)

Bozorgtabar, B., Mahapatra, D., von Teng, H., Pollinger, A., Ebner, L., Thiran, J.P., M.Reyes.: Informative sample generation using class aware generative ad- versarial networks for classiﬁcation of chest xrays. Computer Vision and Image Understanding 184, 57–65 (2019)

work page 2019

[5] [5]

Mahapatra, B.B., Garnavi, R.: Image super-resolution using progressive gen- erative adversarial networks for medical image analysis

D. Mahapatra, B.B., Garnavi, R.: Image super-resolution using progressive gen- erative adversarial networks for medical image analysis. Computerized Medical Imaging and Graphics 71(1), 30–39 (2019)

work page 2019

[6] [6]

al.: Flownet: Learning optical ﬂow with convolu- tional networks

Dosovitskiy, A., Fischer, P., et. al.: Flownet: Learning optical ﬂow with convolu- tional networks. In: In Proc. IEEE ICCV. pp. 2758–2766 (2015)

work page 2015

[7] [7]

In: In Proc

Graziani, M., Andrearczyk, V., Mller, H.: Regression concept vectors for bidirec- tional explanations in histopathology. In: In Proc. MICCAI-iMIMIC. pp. 124–132 (2018)

work page 2018

[8] [8]

In: In Proc

He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. In: In Proc. CVPR (2016)

work page 2016

[9] [9]

Densely Connected Convolutional Networks

Huang, G., Liu, Z., van der Maaten, L., Weinberger, K.: Densely connected con- volutional networks. In: https://arxiv.org/abs/1608.06993, (2016)

work page internal anchor Pith review Pith/arXiv arXiv 2016

[10] [10]

In: In Proc: ACM Intl

Kuang, H., Guthier, B., Saini, M., Mahapatra, D., Saddik, A.E.: A real-time smart assistant for video surveillance through handheld devices. In: In Proc: ACM Intl. Conf. Multimedia. pp. 917–920 (2014)

work page 2014

[11] [11]

IEEE Trans

Li, Z., Mahapatra, D., J.Tielbeek, Stoker, J., van Vliet, L., Vos, F.: Image registra- tion based on autocorrelation of local structure. IEEE Trans. Med. Imaging 35(1), 63–75 (2016)

work page 2016

[12] [12]

In: In Proc: MICCAI workshop on Image Analysis of Human Brain Devel- opment (IAHBD) (2011)

Mahapatra, D.: Neonatal brain mri skull stripping using graph cuts and shape priors. In: In Proc: MICCAI workshop on Image Analysis of Human Brain Devel- opment (IAHBD) (2011)

work page 2011

[13] [13]

In: Proc

Mahapatra, D.: Cardiac lv and rv segmentation using mutual context information. In: Proc. MICCAI-MLMI. pp. 201–209 (2012)

work page 2012

[14] [14]

In: In Proc: SPIE Medical Imaging (2012) 9

Mahapatra, D.: Groupwise registration of dynamic cardiac perfusion images using temporal information and segmentation information. In: In Proc: SPIE Medical Imaging (2012) 9

work page 2012

[15] [15]

In: Proc

Mahapatra, D.: Landmark detection in cardiac mri using learned local image statis- tics. In: Proc. MICCAI-Statistical Atlases and Computational Models of the Heart. Imaging and Modelling Challenges (STACOM). pp. 115–124 (2012)

work page 2012

[16] [16]

Mahapatra, D.: Skull stripping of neonatal brain mri: Using prior shape information with graphcuts. J. Digit. Imaging 25(6), 802–814 (2012)

work page 2012

[17] [17]

Mahapatra, D.: Cardiac image segmentation from cine cardiac mri using graph cuts and shape priors. J. Digit. Imaging 26(4), 721–730 (2013)

work page 2013

[18] [18]

Mahapatra, D.: Cardiac mri segmentation using mutual context information from left and right ventricle. J. Digit. Imaging 26(5), 898–908 (2013)

work page 2013

[19] [19]

In: Proc

Mahapatra, D.: Graph cut based automatic prostate segmentation using learned semantic information. In: Proc. IEEE ISBI. pp. 1304–1307 (2013)

work page 2013

[20] [20]

Mahapatra, D.: Joint segmentation and groupwise registration of cardiac perfusion images using temporal information. J. Digit. Imaging 26(2), 173–182 (2013)

work page 2013

[21] [21]

Mahapatra, D.: Automatic cardiac segmentation using semantic information from random forests. J. Digit. Imaging. 27(6), 794–804 (2014)

work page 2014

[22] [22]

Computer Vision and Image Understanding 151(1), 114–123 (2016)

Mahapatra, D.: Combining multiple expert annotations using semi-supervised learning and graph cuts for medical image segmentation. Computer Vision and Image Understanding 151(1), 114–123 (2016)

work page 2016

[23] [23]

Pattern Recognition 63(1), 700–709 (2017)

Mahapatra, D.: Semi-supervised learning and graph cuts for consensus based med- ical image segmentation. Pattern Recognition 63(1), 700–709 (2017)

work page 2017

[24] [24]

In: In Proc

Mahapatra, D., Bozorgtabar, S., Hewavitahranage, S., Garnavi, R.: Image super resolution using generative adversarial networks and local saliencymaps for retinal image analysis,. In: In Proc. MICCAI. pp. 382–390 (2017)

work page 2017

[25] [25]

In: In Proc

Mahapatra, D., Bozorgtabar, S., Thiran, J.P., Reyes, M.: Eﬃcient active learning for image classiﬁcation and segmentation using a sample selection and conditional generative adversarial network. In: In Proc. MICCAI (2). pp. 580–588 (2018)

work page 2018

[26] [26]

In: In Proc

Mahapatra, D., Buhmann, J.: Obtaining consensus annotations for retinal image segmentation using random forest and graph cuts. In: In Proc. OMIA. pp. 41–48 (2015)

work page 2015

[27] [27]

In: In Proc

Mahapatra, D., Buhmann, J.: Visual saliency based active learning for prostate mri segmentation. In: In Proc. MLMI. pp. 9–16 (2015)

work page 2015

[28] [28]

SPIE Journal of Medical Imaging 3(1) (2016)

Mahapatra, D., Buhmann, J.: Visual saliency based active learning for prostate mri segmentation. SPIE Journal of Medical Imaging 3(1) (2016)

work page 2016

[29] [29]

In: Proc

Mahapatra, D., Buhmann, J.: Automatic cardiac rv segmentation using semantic information with graph cuts. In: Proc. IEEE ISBI. pp. 1094–1097 (2013)

work page 2013

[30] [30]

IEEE Trans

Mahapatra, D., Buhmann, J.: Analyzing training information from random forests for improved image segmentation. IEEE Trans. Imag. Proc. 23(4), 1504–1512 (2014)

work page 2014

[31] [31]

IEEE Trans

Mahapatra, D., Buhmann, J.: Prostate mri segmentation using learned semantic knowledge and graph cuts. IEEE Trans. Biomed. Engg. 61(3), 756–764 (2014)

work page 2014

[32] [32]

In: In Proc

Mahapatra, D., Buhmann, J.: A ﬁeld of experts model for optic cup and disc segmentation from retinal fundus images. In: In Proc. IEEE ISBI. pp. 218–221 (2015)

work page 2015

[33] [33]

In: In Proc

Mahapatra, D., Ge, Z.: Training data independent image registration with gans using transfer learning and segmentation information. In: In Proc. IEEE ISBI (2019)

work page 2019

[34] [34]

In: In Proc

Mahapatra, D., Ge, Z., Sedai, S., Chakravorty., R.: Joint registration and segmen- tation of xray images using generative adversarial networks. In: In Proc. MICCAI- MLMI. pp. 73–80 (2018) 10

work page 2018

[35] [35]

IEEE Journal of Selected Topics in Signal Processing

Mahapatra, D., Gilani, S., Saini., M.: Coherency based spatio-temporal saliency detection for video object segmentation. IEEE Journal of Selected Topics in Signal Processing. 8(3), 454–462 (2014)

work page 2014

[36] [36]

IEEE Trans

Mahapatra, D., J.Tielbeek, Makanyanga, J., Stoker, J., Taylor, S., Vos, F., Buh- mann, J.: Automatic detection and segmentation of crohn’s disease tissues from abdominal mri. IEEE Trans. Med. Imaging 32(12), 1232–1248 (2013)

work page 2013

[37] [37]

In: Proc

Mahapatra, D., J.Tielbeek, Makanyanga, J., Stoker, J., Taylor, S., Vos, F., Buh- mann, J.: Active learning based segmentation of crohn’s disease using principles of visual saliency. In: Proc. IEEE ISBI. pp. 226–229 (2014)

work page 2014

[38] [38]

In: In Proc: MICCAI-ABD (2014)

Mahapatra, D., J.Tielbeek, Makanyanga, J., Stoker, J., Taylor, S., Vos, F., Buh- mann, J.: Combiningmultiple expert annotations using semi-supervised learning and graph cuts for crohns disease segmentation. In: In Proc: MICCAI-ABD (2014)

work page 2014

[39] [39]

Mahapatra, D., J.Tielbeek, Vos, F., Buhmann, J.: A supervised learning approach for crohn’s disease detection using higher order image statistics and a novel shape asymmetry measure. J. Digit. Imaging 26(5), 920–931 (2013)

work page 2013

[40] [40]

In: In Proc

Mahapatra, D., Li, Z., Vos, F., Buhmann, J.: Joint segmentation and groupwise registration of cardiac dce mri using sparse data representations. In: In Proc. IEEE ISBI. pp. 1312–1315 (2015)

work page 2015

[41] [41]

In: IEEE International Conference on Industrial Technology (ICIT)

Mahapatra, D., Routray, A., Mishra, C.: An active snake model for classiﬁcation of extreme emotions. In: IEEE International Conference on Industrial Technology (ICIT). pp. 2195–2199 (2006)

work page 2006

[42] [42]

In: In Proc

Mahapatra, D., Roy, P., Sedai, S., Garnavi, R.: A cnn based neurobiology inspired approach for retinal image quality assessment. In: In Proc. EMBC. pp. 1304–1307 (2016)

work page 2016

[43] [43]

In: In Proc

Mahapatra, D., Roy, P., Sedai, S., Garnavi, R.: Retinal image quality classiﬁcation using saliency maps and cnns. In: In Proc. MICCAI-MLMI. pp. 172–179 (2016)

work page 2016

[44] [44]

In: In 13th International Conference on Biomedical Engineering (2008)

Mahapatra, D., Roy, S., Sun, Y.: Retrieval of mr kidney images by incorporating spatial information in histogram of low level features. In: In 13th International Conference on Biomedical Engineering (2008)

work page 2008

[45] [45]

In: IEEE ICME

Mahapatra, D., Saini, M., Sun, Y.: Illumination invariant tracking in oﬃce envi- ronments using neurobiology-saliency based particle ﬁlter. In: IEEE ICME. pp. 953–956 (2008)

work page 2008

[46] [46]

In: Proc

Mahapatra, D., Sch¨uﬄer, P., Tielbeek, J., Vos, F., Buhmann, J.: Semi-supervised and active learning for automatic segmentation of crohn’s disease. In: Proc. MIC- CAI, Part 2. pp. 214–221 (2013)

work page 2013

[47] [47]

In: Proc

Mahapatra, D., Sun, Y.: Nonrigid registration of dynamic renal MR images using a saliency based MRF model. In: Proc. MICCAI. pp. 771–779 (2008)

work page 2008

[48] [48]

In: Proc

Mahapatra, D., Sun, Y.: Registration of dynamic renal MR images using neurobi- ological model of saliency. In: Proc. ISBI. pp. 1119–1122 (2008)

work page 2008

[49] [49]

In: In 13th International Conference on Biomedical Engi- neering (2008)

Mahapatra, D., Sun, Y.: Using saliency features for graphcut segmentation of per- fusion kidney images. In: In 13th International Conference on Biomedical Engi- neering (2008)

work page 2008

[50] [50]

In: Proc

Mahapatra, D., Sun, Y.: Joint registration and segmentation of dynamic cardiac perfusion images using mrfs. In: Proc. MICCAI. pp. 493–501 (2010)

work page 2010

[51] [51]

In: Proc

Mahapatra, D., Sun., Y.: An mrf framework for joint registration and segmentation of natural and perfusion images. In: Proc. IEEE ICIP. pp. 1709–1712 (2010)

work page 2010

[52] [52]

In: Proc

Mahapatra, D., Sun, Y.: Retrieval of perfusion images using cosegmentation and shape context information. In: Proc. APSIPA Annual Summit and Conference (ASC) (2010) 11

work page 2010

[53] [53]

EURASIP Journal on Image and Video Processing

Mahapatra, D., Sun, Y.: Rigid registration of renal perfusion images using a neu- robiology based visual saliency model. EURASIP Journal on Image and Video Processing. pp. 1–16 (2010)

work page 2010

[54] [54]

In: Proc

Mahapatra, D., Sun, Y.: A saliency based mrf method for the joint registration and segmentation of dynamic renal mr images. In: Proc. ICDIP (2010)

work page 2010

[55] [55]

IEEE Trans

Mahapatra, D., Sun, Y.: Mrf based intensity invariant elastic registration of cardiac perfusion images using saliency information. IEEE Trans. Biomed. Engg. 58(4), 991–1000 (2011)

work page 2011

[56] [56]

In: In Proc: MICCAI

Mahapatra, D., Sun, Y.: Orientation histograms as shape priors for left ventricle segmentation using graph cuts. In: In Proc: MICCAI. pp. 420–427 (2011)

work page 2011

[57] [57]

IEEE Trans

Mahapatra, D., Sun, Y.: Integrating segmentation information for improved mrf- based elastic image registration. IEEE Trans. Imag. Proc. 21(1), 170–183 (2012)

work page 2012

[58] [58]

In: Proc

Mahapatra, D., Tielbeek, J., Buhmann, J., Vos, F.: A supervised learning based ap- proach to detect crohn’s disease in abdominal mr volumes. In: Proc. MICCAI work- shop Computational and Clinical Applications in Abdominal Imaging(MICCAI- ABD). pp. 97–106 (2012)

work page 2012

[59] [59]

In: Proc

Mahapatra, D., Tielbeek, J., Vos, F., ., J.B.: Crohn’s disease tissue segmentation from abdominal mri using semantic information and graph cuts. In: Proc. IEEE ISBI. pp. 358–361 (2013)

work page 2013

[60] [60]

In: Proc

Mahapatra, D., Tielbeek, J., Vos, F., Buhmann, J.: Localizing and segmenting crohn’s disease aﬀected regions in abdominal mri using novel context features. In: Proc. SPIE Medical Imaging (2013)

work page 2013

[61] [61]

In: Proc

Mahapatra, D., Tielbeek, J., Vos, F., Buhmann, J.: Weakly supervised semantic segmentation of crohn’s disease tissues from abdominal mri. In: Proc. IEEE ISBI. pp. 832–835 (2013)

work page 2013

[62] [62]

In: In Proc

Mahapatra, D., Vos, F., Buhmann, J.: Crohn’s disease segmentation from mri using learned image priors. In: In Proc. IEEE ISBI. pp. 625–628 (2015)

work page 2015

[63] [63]

Computer Methods and Programs in Biomedicine 128(1), 75–85 (2016)

Mahapatra, D., Vos, F., Buhmann, J.: Active learning based segmentation of crohns disease from abdominal mri. Computer Methods and Programs in Biomedicine 128(1), 75–85 (2016)

work page 2016

[64] [64]

In: SPIE HVEI

Mahapatra, D., Winkler, S., Yen, S.: Motion saliency outweighs other low-level features while watching videos. In: SPIE HVEI. pp. 1–10 (2008)

work page 2008

[65] [65]

In: In Proc

Pereira, S., Meier, R., Alves, V., Reyes, M., Silva., C.: Automatic brain tumor grading from mri data using convolutional neural networks and quality assessment. In: In Proc. MICCAI-iMIMIC. pp. 106–114 (2018)

work page 2018

[66] [66]

IEEE Trans

Rasti, R., Rabbani, H., Mehri, A., Hajizadeh, F.: Macular oct classiﬁcation using a multi-scale convolutional neural network ensemble. IEEE Trans. Med. Imag. 37(4), 1024–1034 (2018)

work page 2018

[67] [67]

Very Deep Convolutional Networks for Large-Scale Image Recognition

Simonyan, K., Zisserman., A.: Very deep convolutional networks for large-scale image recognition. CoRR abs/1409.1556 (2014)

work page internal anchor Pith review Pith/arXiv arXiv 2014

[68] [68]

In: MIC- CAI

Sokooti, H., de Vos, B., Berendsen, F., Lelieveldt, B., Isgum, I., Staring, M.: Non- rigid image registration using multiscale 3d convolutional neural networks. In: MIC- CAI. pp. 232–239 (2017)

work page 2017

[69] [69]

Journal of Mach

Vincent, P., Larochelle, H., Lajoie, I., Bengio, Y., Manzagol, P.: Stacked denois- ing autoencoders: Learning useful representations in a deep network with a local denoising criterion. Journal of Mach. Learn. Res. 11, 3371–3408 (2010)

work page 2010

[70] [70]

End-to-End Unsupervised Deformable Image Registration with a Convolutional Neural Network

de Vos, B., Berendsen, F., Viergever, M., Staring, M., Isgum, I.: End-to-end un- supervised deformable image registration with a convolutional neural network. In: arXiv preprint arXiv:1704.06065 (2017) 12

work page internal anchor Pith review Pith/arXiv arXiv 2017

[71] [71]

IEEE EMBC

Vos, F.M., Tielbeek, J., Naziroglu, R., Li, Z., Sch¨uﬄer, P., Mahapatra, D., Wiebel, A., Lavini, C., Buhmann, J., Hege, H., Stoker, J., van Vliet, L.: Computational modeling for assessment of IBD: to be or not to be? In: Proc. IEEE EMBC. pp. 3974–3977 (2012)

work page 2012

[72] [72]

In: Proc

Xie, J., Girshick, R., Farhadi, A.: Unsupervised deep embedding for clustering analysis. In: Proc. ICML. pp. 478–487 (2016)

work page 2016

[73] [73]

In: Proc

Zhou, B., Khosla, A., Lapedriza, A., Oliva, A., Torralba, A.: Learning deep features for discriminative localization. In: Proc. CVPR. pp. 2921–2929 (2016)

work page 2016

[74] [74]

In: In Proc

Zilly, J., Buhmann, J., Mahapatra, D.: Boosting convolutional ﬁlters with entropy sampling for optic cup and disc image segmentation from fundus images. In: In Proc. MLMI. pp. 136–143 (2015)

work page 2015

[75] [75]

In Press Computerized Medical Imaging and Graphics 55(1), 28–41 (2017)

Zilly, J., Buhmann, J., Mahapatra, D.: Glaucoma detection using entropy sampling and ensemble learning for automatic optic cup and disc segmentation. In Press Computerized Medical Imaging and Graphics 55(1), 28–41 (2017)

work page 2017