Erwan Zerhouni

Bio: Erwan Zerhouni is an academic researcher from IBM. The author has contributed to research in topics: Autoencoder & Feature extraction. The author has an hindex of 4, co-authored 7 publications receiving 170 citations.

Wide residual networks for mitosis detection

Abstract: One of the most important prognostic markers to assess proliferation activity of breast tumors is estimating the number of mitotic figures in H&E stained tissue. We propose the use of a recently published convolutional neural network architecture, Wide Residual Networks, for mitosis detection in breast histology images. The model is trained to classify each pixel of on an image using as context a patch centered on the pixel. We apply post-processing on the network output in order to filter out noise and select true mitosis. Finally, we combine the output of several networks using majority vote. Our approach ranked 2nd in the MICCAI TUPAC 2016 competition for mitosis detection, outperforming most other contestants by a significant margin.

Simultaneous feature extraction and dictionary learning using deep learning architectures for characterization of images of heterogeneous tissue samples

Abstract: Apparatus, methods, and computer-readable media are provided for simultaneous feature extraction and dictionary learning from heterogeneous tissue images, without the need of prior local labeling. A convolutional autoencoder is adapted and enhanced to jointly learn a feature extraction algorithm and a dictionary of representative atoms. While training the autoencoder an image patch is tiled in sub-patches and only the highest activation value per sub-patch is kept. Thus, only a subset of spatially constrained values per patch is used for reconstruction. The deconvolutional filters are the dictionary elements, and only a deconvolution layer is used for these elements. Embodiments described herein may be provided for use in models for representing local tissue heterogeneity for better disease progression understanding and thus treating, diagnosing, and/or predicting the occurrence (e.g., recurrence) of one or more medical conditions such as, for example, cancer or other types of disease.

A computational framework for disease grading using protein signatures

Abstract: In this paper, we propose a novel framework for computational disease stratification based on protein expression tissue images. We extract cellular staining response using color information and create a graph based on morphological features and their spatial distance. This graph is collapsed using a learned dictionary. We then compute the commute time matrix and use it as unique signature per protein and disease grade. We combine protein-based signatures using SVM with an Multiple Kernel Learning approach. We test the proposed framework on a prostate cancer tissue dataset and demonstrate the efficacy of the derived protein signatures for both disease stratification and quantification of the relative importance of each protein.

Heterogeneity characterization of immunohistochemistry stained tissue using convolutional autoencoder

Abstract: The focus of this paper is to illustrate how computational image processing and machine learning can help address two of the challenges of histological image analysis, namely, the cellular heterogeneity, and the imprecise labeling. We propose an unsupervised method of generating representative image signatures based on an autoencoder architecture which reduces the dependency on labels that tend to be imprecise and tedious to get. We have modified and enhanced the architecture to simultaneously produce representative image features as well as perform dictionary learning on these features to enable robust characterization of the cellular phenotypes. We integrate the extracted features in a disease grading framework, test it in prostate tissues immunostained for different protein visualization and show significant improvement in terms of grading accuracy compared to alternative supervised feature-extraction methods.

Pyramid Vision Transformer: A Versatile Backbone for Dense Prediction without Convolutions

Abstract: Although using convolutional neural networks (CNNs) as backbones achieves great successes in computer vision, this work investigates a simple backbone network useful for many dense prediction tasks without convolutions. Unlike the recently-proposed Transformer model (e.g., ViT) that is specially designed for image classification, we propose Pyramid Vision Transformer~(PVT), which overcomes the difficulties of porting Transformer to various dense prediction tasks. PVT has several merits compared to prior arts. (1) Different from ViT that typically has low-resolution outputs and high computational and memory cost, PVT can be not only trained on dense partitions of the image to achieve high output resolution, which is important for dense predictions but also using a progressive shrinking pyramid to reduce computations of large feature maps. (2) PVT inherits the advantages from both CNN and Transformer, making it a unified backbone in various vision tasks without convolutions by simply replacing CNN backbones. (3) We validate PVT by conducting extensive experiments, showing that it boosts the performance of many downstream tasks, e.g., object detection, semantic, and instance segmentation. For example, with a comparable number of parameters, RetinaNet+PVT achieves 40.4 AP on the COCO dataset, surpassing RetinNet+ResNet50 (36.3 AP) by 4.1 absolute AP. We hope PVT could serve as an alternative and useful backbone for pixel-level predictions and facilitate future researches. Code is available at this https URL.

Prognostic factors in breast cancer

Abstract: Prognostic factors are clinical and pathological features that give information in estimating the likely clinical outcome of an individual suffering from cancer. The author gives a short review of the most important prognostic factors in breast cancer. 376 breast cancer cases of a ten year interval in a county hospital are summarized. Traditional clinico-pathological parameters i.e. TNM and steroid receptor status are discussed. The more common karyotipic, oncogene and tumor suppressor gene alterations are outlined in the study. Methods for their detection are presented and their value in prognostication is reviewed. Emphasis was laid on steroid receptors, c-erpB-2, p53 and bcl-2 alterations. Genes responsible for heritable forms of increased breast cancer risk are briefly reviewed.

WILDS: A Benchmark of in-the-Wild Distribution Shifts

Abstract: Distribution shifts -- where the training distribution differs from the test distribution -- can substantially degrade the accuracy of machine learning (ML) systems deployed in the wild. Despite their ubiquity, these real-world distribution shifts are under-represented in the datasets widely used in the ML community today. To address this gap, we present WILDS, a curated collection of 8 benchmark datasets that reflect a diverse range of distribution shifts which naturally arise in real-world applications, such as shifts across hospitals for tumor identification; across camera traps for wildlife monitoring; and across time and location in satellite imaging and poverty mapping. On each dataset, we show that standard training results in substantially lower out-of-distribution than in-distribution performance, and that this gap remains even with models trained by existing methods for handling distribution shifts. This underscores the need for new training methods that produce models which are more robust to the types of distribution shifts that arise in practice. To facilitate method development, we provide an open-source package that automates dataset loading, contains default model architectures and hyperparameters, and standardizes evaluations. Code and leaderboards are available at this https URL.

Automated Gleason grading of prostate cancer tissue microarrays via deep learning

Abstract: The Gleason grading system remains the most powerful prognostic predictor for patients with prostate cancer since the 1960s. Its application requires highly-trained pathologists, is tedious and yet suffers from limited inter-pathologist reproducibility, especially for the intermediate Gleason score 7. Automated annotation procedures constitute a viable solution to remedy these limitations. In this study, we present a deep learning approach for automated Gleason grading of prostate cancer tissue microarrays with Hematoxylin and Eosin (H&E) staining. Our system was trained using detailed Gleason annotations on a discovery cohort of 641 patients and was then evaluated on an independent test cohort of 245 patients annotated by two pathologists. On the test cohort, the inter-annotator agreements between the model and each pathologist, quantified via Cohen's quadratic kappa statistic, were 0.75 and 0.71 respectively, comparable with the inter-pathologist agreement (kappa = 0.71). Furthermore, the model's Gleason score assignments achieved pathology expert-level stratification of patients into prognostically distinct groups, on the basis of disease-specific survival data available for the test cohort. Overall, our study shows promising results regarding the applicability of deep learning-based solutions towards more objective and reproducible prostate cancer grading, especially for cases with heterogeneous Gleason patterns.