Other affiliations: École des ponts ParisTech, Université Paris-Saclay, École Centrale Paris ...read more
Bio: Stavros Tsogkas is an academic researcher from University of Toronto. The author has contributed to research in topics: Convolutional neural network & Segmentation. The author has an hindex of 16, co-authored 33 publications receiving 764 citations. Previous affiliations of Stavros Tsogkas include École des ponts ParisTech & Université Paris-Saclay.
23 Jun 2014
TL;DR: A dataset of 7, 413 airplanes annotated in detail with parts and their attributes is introduced, leveraging images donated by airplane spotters and crowd-sourcing both the design and collection of the detailed annotations to provide insights that should help researchers interested in designing fine-grained datasets for other basic level categories.
Abstract: We study the problem of understanding objects in detail, intended as recognizing a wide array of fine-grained object attributes. To this end, we introduce a dataset of 7, 413 airplanes annotated in detail with parts and their attributes, leveraging images donated by airplane spotters and crowdsourcing both the design and collection of the detailed annotations. We provide a number of insights that should help researchers interested in designing fine-grained datasets for other basic level categories. We show that the collected data can be used to study the relation between part detection and attribute prediction by diagnosing the performance of classifiers that pool information from different parts of an object. We note that the prediction of certain attributes can benefit substantially from accurate part detection. We also show that, differently from previous results in object detection, employing a large number of part templates can improve detection accuracy at the expenses of detection speed. We finally propose a coarse-to-fine approach to speed up detection through a hierarchical cascade algorithm.
••07 Oct 2012
TL;DR: This work creates and makes publicly available a ground-truth dataset for symmetry detection in natural images, and uses supervised learning to learn how to combine these cues, and employs MIL to accommodate the unknown scale and orientation of the symmetric structures.
Abstract: In this work we propose a learning-based approach to symmetry detection in natural images. We focus on ribbon-like structures, i.e. contours marking local and approximate reflection symmetry and make three contributions to improve their detection. First, we create and make publicly available a ground-truth dataset for this task by building on the Berkeley Segmentation Dataset. Second, we extract features representing multiple complementary cues, such as grayscale structure, color, texture, and spectral clustering information. Third, we use supervised learning to learn how to combine these cues, and employ MIL to accommodate the unknown scale and orientation of the symmetric structures. We systematically evaluate the performance contribution of each individual component in our pipeline, and demonstrate that overall we consistently improve upon results obtained using existing alternatives.
••13 Apr 2016
TL;DR: In this article, a Fully-Convolutional Neural Network (F-CNN) was used to segment sub-cortical structures of the human brain in Magnetic Resonance (MR) image data.
Abstract: In this paper we propose a deep learning approach for segmenting sub-cortical structures of the human brain in Magnetic Resonance (MR) image data. We draw inspiration from a state-of-the-art Fully-Convolutional Neural Network (F-CNN) architecture for semantic segmentation of objects in natural images, and adapt it to our task. Unlike previous CNN-based methods that operate on image patches, our model is applied on a full blown 2D image, without any alignment or registration steps at testing time. We further improve segmentation results by interpreting the CNN output as potentials of a Markov Random Field (MRF), whose topology corresponds to a volumetric grid. Alpha-expansion is used to perform approximate inference imposing spatial volumetric homogeneity to the CNN priors. We compare the performance of the proposed pipeline with a similar system using Random Forest-based priors, as well as state-of-art segmentation algorithms, and show promising results on two different brain MRI datasets.
TL;DR: A state-of-the-art semantic segmentation system is adapted to this task, and it is shown that a combination of a fully-convolutional Deep CNN system coupled with Dense CRF labelling provides excellent results for a broad range of object categories.
Abstract: In this work we address the task of segmenting an object into its parts, or semantic part segmentation. We start by adapting a state-of-the-art semantic segmentation system to this task, and show that a combination of a fully-convolutional Deep CNN system coupled with Dense CRF labelling provides excellent results for a broad range of object categories. Still, this approach remains agnostic to high-level constraints between object parts. We introduce such prior information by means of the Restricted Boltzmann Machine, adapted to our task and train our model in an discriminative fashion, as a hidden CRF, demonstrating that prior information can yield additional improvements. We also investigate the performance of our approach ``in the wild'', without information concerning the objects' bounding boxes, using an object detector to guide a multi-scale segmentation scheme. We evaluate the performance of our approach on the Penn-Fudan and LFW datasets for the tasks of pedestrian parsing and face labelling respectively. We show superior performance with respect to competitive methods that have been extensively engineered on these benchmarks, as well as realistic qualitative results on part segmentation, even for occluded or deformable objects. We also provide quantitative and extensive qualitative results on three classes from the PASCAL Parts dataset. Finally, we show that our multi-scale segmentation scheme can boost accuracy, recovering segmentations for finer parts.
TL;DR: The AIM 2020 challenge on efficient single image super-resolution was to super-resolve an input image with a magnification factor x4 based on a set of prior examples of low and corresponding high resolution images with focus on the proposed solutions and results.
Abstract: This paper reviews the AIM 2020 challenge on efficient single image super-resolution with focus on the proposed solutions and results. The challenge task was to super-resolve an input image with a magnification factor x4 based on a set of prior examples of low and corresponding high resolution images. The goal is to devise a network that reduces one or several aspects such as runtime, parameter count, FLOPs, activations, and memory consumption while at least maintaining PSNR of MSRResNet. The track had 150 registered participants, and 25 teams submitted the final results. They gauge the state-of-the-art in efficient single image super-resolution.
TL;DR: This paper reviews the major deep learning concepts pertinent to medical image analysis and summarizes over 300 contributions to the field, most of which appeared in the last year, to survey the use of deep learning for image classification, object detection, segmentation, registration, and other tasks.
TL;DR: This work brings together methods from DCNNs and probabilistic graphical models for addressing the task of pixel-level classification by combining the responses at the final DCNN layer with a fully connected Conditional Random Field (CRF).
Abstract: Deep Convolutional Neural Networks (DCNNs) have recently shown state of the art performance in high level vision tasks, such as image classification and object detection. This work brings together methods from DCNNs and probabilistic graphical models for addressing the task of pixel-level classification (also called "semantic image segmentation"). We show that responses at the final layer of DCNNs are not sufficiently localized for accurate object segmentation. This is due to the very invariance properties that make DCNNs good for high level tasks. We overcome this poor localization property of deep networks by combining the responses at the final DCNN layer with a fully connected Conditional Random Field (CRF). Qualitatively, our "DeepLab" system is able to localize segment boundaries at a level of accuracy which is beyond previous methods. Quantitatively, our method sets the new state-of-art at the PASCAL VOC-2012 semantic image segmentation task, reaching 71.6% IOU accuracy in the test set. We show how these results can be obtained efficiently: Careful network re-purposing and a novel application of the 'hole' algorithm from the wavelet community allow dense computation of neural net responses at 8 frames per second on a modern GPU.
TL;DR: This paper details the improvements of CNN on different aspects, including layer design, activation function, loss function, regularization, optimization and fast computation, and introduces various applications of convolutional neural networks in computer vision, speech and natural language processing.
Abstract: In the last few years, deep learning has led to very good performance on a variety of problems, such as visual recognition, speech recognition and natural language processing. Among different types of deep neural networks, convolutional neural networks have been most extensively studied. Leveraging on the rapid growth in the amount of the annotated data and the great improvements in the strengths of graphics processor units, the research on convolutional neural networks has been emerged swiftly and achieved state-of-the-art results on various tasks. In this paper, we provide a broad survey of the recent advances in convolutional neural networks. We detailize the improvements of CNN on different aspects, including layer design, activation function, loss function, regularization, optimization and fast computation. Besides, we also introduce various applications of convolutional neural networks in computer vision, speech and natural language processing.
TL;DR: This paper proposed bilinear models, which consists of two feature extractors whose outputs are multiplied using outer product at each location of the image and pooled to obtain an image descriptor, which can model local pairwise feature interactions in a translationally invariant manner.
Abstract: We propose bilinear models, a recognition architecture that consists of two feature extractors whose outputs are multiplied using outer product at each location of the image and pooled to obtain an image descriptor. This architecture can model local pairwise feature interactions in a translationally invariant manner which is particularly useful for fine-grained categorization. It also generalizes various orderless texture descriptors such as the Fisher vector, VLAD and O2P. We present experiments with bilinear models where the feature extractors are based on convolutional neural networks. The bilinear form simplifies gradient computation and allows end-to-end training of both networks using image labels only. Using networks initialized from the ImageNet dataset followed by domain specific fine-tuning we obtain 84.1% accuracy of the CUB-200-2011 dataset requiring only category labels at training time. We present experiments and visualizations that analyze the effects of fine-tuning and the choice two networks on the speed and accuracy of the models. Results show that the architecture compares favorably to the existing state of the art on a number of fine-grained datasets while being substantially simpler and easier to train. Moreover, our most accurate model is fairly efficient running at 8 frames/sec on a NVIDIA Tesla K40 GPU. The source code for the complete system will be made available at this http URL