{"forum": "B1geHYXgx4", "submission_url": "https://openreview.net/forum?id=B1geHYXgx4", "submission_content": {"title": "End-to-End Image-to-Tree for Vasculature Modeling", "authors": ["Manish Sharma", "Matthew C. H. Lee", "James Batten", "Michiel Schaap", "Ben Glocker"], "authorids": ["skmanish@google.com", "mathewlee13@imperial.ac.uk", "j.batten@imperial.ac.uk", "mschaap@heartflow.com", "b.glocker@imperial.ac.uk"], "keywords": ["Vasculature", "tree extraction", "retinal vessels", "diabetic retinopathy", "visualization"], "abstract": "Imaging can be used to capture detailed information about complex anatomical structures such as vessel trees. This can help to detect disease such as stenosis (blockages) which is important for diagnosis and clinical decision making. Current approaches for extracting vasculature from images involve generating binary segmentation maps followed by further processing. However, these binary maps may be sub-optimal, implicit representations of the underlying geometry while trees seem a more natural way of describing vasculature. In this work, we propose a novel image-to-tree approach, which is an end-to-end system for extracting explicit tree representations of vasculature from biomedical scans. We designed a moving patch algorithm that utilizes a U-Net component for predicting individual tree nodes. The methodology is presented for both synthetically generated tree images and publicly available Digital Retinal Vessel Extraction dataset (DRIVE). Using vascular tree construction, we discuss applications to thickness estimation in diabetic retinopathy prediction, and explore insights from visualizing these trees.", "pdf": "/pdf/5c6fb1ff59f49dfbb3d92255b77d92520d8cf370.pdf", "code of conduct": "I have read and accept the code of conduct.", "paperhash": "sharma|endtoend_imagetotree_for_vasculature_modeling"}, "submission_cdate": 1544726840120, "submission_tcdate": 1544726840120, "submission_tmdate": 1545069832717, "submission_ddate": null, "review_id": ["Bklp_zeAM4", "ryg6F0-7QE", "BkeA9-22mV"], "review_url": ["https://openreview.net/forum?id=B1geHYXgx4&noteId=Bklp_zeAM4", "https://openreview.net/forum?id=B1geHYXgx4&noteId=ryg6F0-7QE", "https://openreview.net/forum?id=B1geHYXgx4&noteId=BkeA9-22mV"], "review_cdate": [1547727477281, 1548062340666, 1548693910084], "review_tcdate": [1547727477281, 1548062340666, 1548693910084], "review_tmdate": [1549879453231, 1548856737679, 1548856699690], "review_readers": [["everyone"], ["everyone"], ["everyone"]], "review_writers": [["MIDL.io/2019/Conference/Paper88/AnonReviewer2"], ["MIDL.io/2019/Conference/Paper88/AnonReviewer1"], ["MIDL.io/2019/Conference/Paper88/AnonReviewer3"]], "review_reply_count": [{"replyCount": 0}, {"replyCount": 0}, {"replyCount": 0}], "review_replyto": ["B1geHYXgx4", "B1geHYXgx4", "B1geHYXgx4"], "review_content": [{"pros": "- The authors present an end-to-end approach that allows to retrieve vessel trees straight from images, segmentations or skeletonizations, based on a patch-guided U-Net.\n\n- Different deep neural network architectures are explored on a series of artificial images generated with two alternative, novel hand-crafted approaches. The observations indicate that the best performing algorithm is the U-Net.\n\n- This final model is studied in the context of vasculature characterization in fundus images. This is an important step in several clinical studies that are focused on analyzing correlation between vascular characteristics and disease progression.\n\n- Two techniques for visualizing graphs are applied on the outputs of the networks. To the best of my knowledge, this is the first time that these algorithms are applied to visualize retinal vascular trees.\n\n- The paper is written in an excellent style. It is easy to follow, the explanations are simple and therefore straightforward, and the experimental setup is well designed. The reader can certainly follow each experiments step by step, and comprehensible understand the contribution of each components of the proposal.\n\n- Despite the fact that the deep learning contribution is not too significant, this method can certainly contribute to the field of ophthalmic image analysis, specially in clinical studies where the anatomical vessel properties are analyzed.", "cons": "- Authors refer to their approach as \"end-to-end image-to-tree\", but when evaluated on real images the results are not as good as when using segmentations or skeletonizations of the vessels as inputs. This is an important issue and I think that authors should take that result into consideration and modify the claims (and perhaps the title) accordingly. Provided that these modificiations are done, I believe that the article could be certainly accepted. Current performance of vessel segmentation algorithms is close to the one of human observers doing the task manually, so using a segmentation as input would not be really a problem.\n\n- It it not sufficiently emphasize that the methods for synthesizing vascular trees are novel and were not explored before. \n\n- The algorithm requires a starting point to extract the vascular graph. This position is by definition the central, top pixel in synthetic images. However, it is not clear which point is used when working on retinal images. This is also an important thing to consider. Using a single vessel from the optic disc is usually not enough, as some images might show more than one vessel spreading from this region. In [1-4], all the models solve the issue by taking root nodes in the intersection of the optic disc border and vessels. Did you follow a similar idea?\n\n- Is the model based on segmentations (not in skeletonizations) able to solve vessel crossings such as the one illustrated in Fig. 4 (c), bottom? Usually the skeletonization algorithms introduce a small piece of vessel there due to the overlap between vessels. If the proposed method is able to overcome that issue, then it might have really good implications in many applications, including blood flow simulation [1-4], where these ambiguities introduce false branching points that significantly affect the results.\n\n- Some other minor suggestions:\n\n--> It should be clarified in the introduction that Fraz et al. survey is focused only on retinal images and not in blood vessel segmentation in general.\n--> In lines 8 and 9 of the introduction, there is a repetition (\"biomedical scans\").\n--> Although it is clear that the estimated vessel width is correlated with the manual annotations (Fig. 7 (b)), it would be interesting to complement those results with the R^2 value of a linear regression model and a Pearson correlation coefficient.\n\n\nReferences:\n\n[1] Liu, D., Wood, N. B., Xu, X. Y., Witt, N., Hughes, A. D., & Thom, S. A. (2009). Image-based blood flow simulation in the retinal circulation. In 4th European Conference of the International Federation for Medical and Biological Engineering (pp. 1963-1966). Springer, Berlin, Heidelberg.\n\n[2] Malek, J., Azar, A. T., Nasralli, B., Tekari, M., Kamoun, H., & Tourki, R. (2015). Computational analysis of blood flow in the retinal arteries and veins using fundus image. Computers & Mathematics with Applications, 69(2), 101-116.\n\n[3] Caliv\u00e1, F., Leontidis, G., Chudzik, P., Hunter, A., Antiga, L., & Al-Diri, B. (2017). Hemodynamics in the retinal vasculature during the progression of diabetic retinopathy. Journal for modeling in Ophthalmology, 1(4), 6-15.\n\n[4] Orlando J.I., Barbosa Breda J., van Keer K., Blaschko M.B., Blanco P.J., Bulant C.A. (2018) Towards a Glaucoma Risk Index Based on Simulated Hemodynamics from Fundus Images. In: Frangi A., Schnabel J., Davatzikos C., Alberola-L\u00f3pez C., Fichtinger G. (eds) Medical Image Computing and Computer Assisted Intervention \u2013 MICCAI 2018. MICCAI 2018. Lecture Notes in Computer Science, vol 11071. Springer, Cham", "rating": "2: reject", "confidence": "3: The reviewer is absolutely certain that the evaluation is correct and very familiar with the relevant literature"}, {"pros": "-\tThe authors compare RNN approaches with iterative CNN approaches, leading to the insight that for this task iterative application of a CNN performs much better than an RNN.\n-\tA synthetic vessel data set is generated to develop the method. This is a potentially useful contribution for development of such methods, but should also be put into context with similar existing works (e.g. https://doi.org/10.1016/j.compmedimag.2010.06.002). \n", "cons": "-\tI don't think the proposed method is an \u2018image-to-tree\u2019 method. It actually performs an iterative segmentation of the voxels that make up the vessel centerlines in the image, i.e. deep learning-based region growing. The result is a segmentation mask which is in principle simlar to a thinned version of a segmentation of the retinal vessels. A similar result might be obtained by first obtaining a binary retinal vessel segmentation (for which many DL methods have been proposed, e.g. https://doi.org/10.1007/978-3-319-46723-8_17, https://doi.org/10.1109/TMI.2016.2546227) and then applying a conventional morphological thinning operation.  The obtained segmentation does not contain information about the topology of the vessels (e.g. separation of veins and arteries, branching points, individual segments) that would facilitate more advanced tree analysis (e.g. https://doi.org/10.1109/TPAMI.2012.265). \n-\tI don\u2019t agree with the authors that this is an \u2018end-to-end\u2019 method. The best performing method is found to be an approach in which a CNN iteratively provides a prediction of the most likely prediction to a tracker. This is not end-to-end, as the CNN is used many times to provide a prediction.  \n-\tReferences to related work are missing. Vessel segmentation/tracking has a long history, see e.g. the review by Lesage et al. (https://doi.org/10.1016/j.media.2009.07.011), multi-orientation tracking by Friman et al. (https://doi.org/10.1016/j.media.2009.12.003), work by Bekkers et al. (https://doi.org/10.1007/s10851-013-0488-6). DL methods for vessel tracking include simultaneous orientation classification and radius prediction (Wolterink et al. https://doi.org/10.1016/j.media.2018.10.005) and LSTM-based methods (Poulin et al. https://doi.org/10.1007/978-3-319-66182-7_62). \n-\tThe method is evaluated on the DRIVE data set, which is an old data set consisting of relatively small and old fundus images. It would be interesting to see how the method fares on larger images such as the High-Resolution Fundus (HRF) image data set or the REVIEW data set. In addition, as many vessels are visualized with 3D imaging it would be good to evaluate the method on a 3D data set, e.g. http://coronary.bigr.nl/centerlines/ or http://image.diku.dk/exact/). \n-\tIt is unclear how the method deals with vessels running in parallel. Based on Fig. 4 and the description in the text, the method would be trained to jump from one vessel to the other. This would be highly undesirable when differentiating between e.g. arteries and veins. In fact, the example results in Fig. 10 show a lot of cyclic structures, which indicates that the tracker connects arteries and veins.\n-\tAdditional constructive feedback:\no\tFigs. 3 and 6 are a bit unconventional. It would have been nicer to show precision and recall in one plot (as you actually do in Fig. 3B) and use isolines to indicate Dice/F1 scores in those images.\no\tThe deep convolutional network (DCN) is not described anywhere.\no\tThere is no description or discussion of the results shown in Fig. 3, while these may actually be the main insight of the paper.", "rating": "1: strong reject", "confidence": "3: The reviewer is absolutely certain that the evaluation is correct and very familiar with the relevant literature"}, {"pros": "Summary: \n\nIn this work, a patch based approach to obtaining trees from image data is presented. Neural networks are used to train patch based predictors that predict nodes, which are then used to successively build trees of interest. Different neural networks are evaluated on synthetically generated data and U-net based patch predictor is finalised. This model is then evaluated on DRIVE data, comprising colour retinal images. Further, an updated U-net based regressor is used to predict vessel width. The preliminary evaluation presented is inconclusive as no relevant comparing methods are presented.\n\nPros:\n- The primary motivation of the work is interesting: to go directly from images to trees instead of binary mask based segmentation\n- Use of neural networks to predict possible nodes in trees\n- The visualisation related work in Section 4 can be interesting. Perhaps it warrants a stand-alone short paper submission, as it does not blend fluently with the rest of the paper.\n", "cons": "- With the larger objective of going from image-to-tree, the presented evaluations appear incomplete. For instance, the evaluation metrics are computed on \"binary masks of tree generated\" (Sec 2.2). This would, in my opinion, contradict the primary objective of bypassing the binary segmentation step. While not straightforward, there are works on tree-space statistics that can be used to perform evaluations directly on trees (for example in [1]). This will considerably strengthen the work by aligning it with its primary objective. \n\n- In choosing the neural networks, it is mentioned that the sequence-less models work better than the sequence-based ones, without a discussion. One would hope that in a recurrent setting, there is more information for making improved node predictions. So, it is surprising.\n\n- This brings me to my next question: Instead of using sequence-less neural networks to predict individual nodes on small patches, why not train the networks like U-net to predict all possible nodes on the entire image? \n\n- Section 3.1 is ambiguous. It describes three levels of vascular tree construction in \"increasing order of difficulty\". Do the authors see each of these tasks as going from image-to-tree? Because the evaluations in Figure 6 seem to indicate this. Obtaining trees from retinal images is what is most interesting, and this seemed to be the motivation presented earlier. Given a segmentation map, obtaining a skeleton and then a tree from it is not as interesting. As a result, the comparisons presented in Figure 6 do not tell much. It is not surprising that the skeleton-to-tree is better, as the segmentation task is already solved.\n\n[1] https://di.ku.dk/forskning/Publikationer/tekniske_rapporter/2011/techrep_trees_Aasa_Feragen.pdf", "rating": "2: reject", "confidence": "2: The reviewer is fairly confident that the evaluation is correct"}], "comment_id": ["SkgnWvLhV4", "Bkxmr_LnVV", "HyxONKUn4N"], "comment_cdate": [1549719299565, 1549719611090, 1549719855641], "comment_tcdate": [1549719299565, 1549719611090, 1549719855641], "comment_tmdate": [1555945997328, 1555945997106, 1555945996889], "comment_readers": [["everyone"], ["everyone"], ["everyone"]], "comment_writers": [["MIDL.io/2019/Conference/Paper88/Authors", "MIDL.io/2019/Conference"], ["MIDL.io/2019/Conference/Paper88/Authors", "MIDL.io/2019/Conference"], ["MIDL.io/2019/Conference/Paper88/Authors", "MIDL.io/2019/Conference"]], "comment_reply_content": [{"replyCount": 0}, {"replyCount": 0}, {"replyCount": 0}], "comment_content": [{"title": "Thank you for the review", "comment": "We really appreciate the time you took for writing this elaborate feedback. \nWe agree with you that our paper lacks some rigorous analysis using tree based statistics. At some places like sec 3.1 the explanation does get ambiguous. \nWe hope to make these things more clear in our next submission.\n\nThanks again!"}, {"title": "Thank you for the review", "comment": "We agree with you on most points that you wrote above. Our work needs to be positioned in the context of slightly more previous work and needs more analysis on 3D or high resolution data.\nWe will keep these in mind while preparing for our next submission. Thanks again!"}, {"title": "Thank you for the review ", "comment": "Thanks for the elaborate feedback. We found that there are certain flaws in the way our paper is structured and we hope to work on the feedback given by you and other reviewers to restructure the content again. "}], "comment_replyto": ["BkeA9-22mV", "ryg6F0-7QE", "Bklp_zeAM4"], "comment_url": ["https://openreview.net/forum?id=B1geHYXgx4&noteId=SkgnWvLhV4", "https://openreview.net/forum?id=B1geHYXgx4&noteId=Bkxmr_LnVV", "https://openreview.net/forum?id=B1geHYXgx4&noteId=HyxONKUn4N"], "meta_review_cdate": 1551356552508, "meta_review_tcdate": 1551356552508, "meta_review_tmdate": 1551703105383, "meta_review_ddate ": null, "meta_review_title": "Acceptance Decision", "meta_review_metareview": "Given the facts that the reviews led to a clear overall recommendation (2 reject, 1 strong reject), and that authors expressed their agreement with many points that were brought up in the reviews, I believe the decision to reject this manuscript from MIDL 2019 is beyond dispute. However, I would like to echo a point that came up repeatedly in the reviews: Generating vessel trees from images is an important challenge in medical image analysis, and approaches to learn this task end-to-end will be a valuable addition to the literature. Therefore, authors should be encouraged to continue their efforts and to resubmit next year, or to another venue.", "meta_review_readers": ["everyone"], "meta_review_writers": ["MIDL.io/2019/Conference"], "meta_review_reply_count": {"replyCount": 0}, "meta_review_url": ["https://openreview.net/forum?id=B1geHYXgx4&noteId=SkegcMUB8N"], "decision": "Reject"}