Deeper neural networks are more difficult to train. We present a residual learning framework to ease the training of networks that are substantially deeper than those used previously. We explicitly reformulate the layers as learning residual functions with reference to the layer inputs, instead of learning unreferenced functions. We provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth. On the ImageNet dataset we evaluate residual nets with a depth of up to 152 layers—8× deeper than VGG nets [40] but still having lower complexity. An ensemble of these residual nets achieves 3.57% error on the ImageNet test set. This result won the 1st place on the ILSVRC 2015 classification task. We also present analysis on CIFAR-10 with 100 and 1000 layers. The depth of representations is of central importance for many visual recognition tasks. Solely due to our extremely deep representations, we obtain a 28% relative improvement on the COCO object detection dataset. Deep residual nets are foundations of our submissions to ILSVRC & COCO 2015 competitions1, where we also won the 1st places on the tasks of ImageNet detection, ImageNet localization, COCO detection, and COCO segmentation.

/pdf/deep-residual-learning-for-image-recognition-b75jg61qrq.pdf

Deep Residual Learning for Image Recognition

In this work we investigate the effect of the convolutional network depth on its accuracy in the large-scale image recognition setting. Our main contribution is a thorough evaluation of networks of increasing depth using an architecture with very small (3x3) convolution filters, which shows that a significant improvement on the prior-art configurations can be achieved by pushing the depth to 16-19 weight layers. These findings were the basis of our ImageNet Challenge 2014 submission, where our team secured the first and the second places in the localisation and classification tracks respectively. We also show that our representations generalise well to other datasets, where they achieve state-of-the-art results. We have made our two best-performing ConvNet models publicly available to facilitate further research on the use of deep visual representations in computer vision.

Very Deep Convolutional Networks for Large-Scale Image Recognition

Deeper neural networks are more difficult to train. We present a residual learning framework to ease the training of networks that are substantially deeper than those used previously. We explicitly reformulate the layers as learning residual functions with reference to the layer inputs, instead of learning unreferenced functions. We provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth. On the ImageNet dataset we evaluate residual nets with a depth of up to 152 layers---8x deeper than VGG nets but still having lower complexity. An ensemble of these residual nets achieves 3.57% error on the ImageNet test set. This result won the 1st place on the ILSVRC 2015 classification task. We also present analysis on CIFAR-10 with 100 and 1000 layers. 
The depth of representations is of central importance for many visual recognition tasks. Solely due to our extremely deep representations, we obtain a 28% relative improvement on the COCO object detection dataset. Deep residual nets are foundations of our submissions to ILSVRC & COCO 2015 competitions, where we also won the 1st places on the tasks of ImageNet detection, ImageNet localization, COCO detection, and COCO segmentation.

We propose a deep convolutional neural network architecture codenamed Inception that achieves the new state of the art for classification and detection in the ImageNet Large-Scale Visual Recognition Challenge 2014 (ILSVRC14). The main hallmark of this architecture is the improved utilization of the computing resources inside the network. By a carefully crafted design, we increased the depth and width of the network while keeping the computational budget constant. To optimize quality, the architectural decisions were based on the Hebbian principle and the intuition of multi-scale processing. One particular incarnation used in our submission for ILSVRC14 is called GoogLeNet, a 22 layers deep network, the quality of which is assessed in the context of classification and detection.

/pdf/going-deeper-with-convolutions-1yobw2o2ds.pdf

Going deeper with convolutions

Classifying facial expressions into different categories requires capturing regional distortions of facial landmarks. We believe that second-order statistics such as covariance is better able to capture such distortions in regional facial features. In this work, we explore the benefits of using a manifold network structure for covariance pooling to improve facial expression recognition. In particular, we first employ such kind of manifold networks in conjunction with traditional convolutional networks for spatial pooling within individual image feature maps in an end-to-end deep learning manner. By doing so, we are able to achieve a recognition accuracy of 58.14% on the validation set of Static Facial Expressions in the Wild (SFEW2.0) and 87.0% on the validation set of Real-World Affective Faces (RAF) Database1. Both of these results are the best results we are aware of. Besides, we leverage covariance pooling to capture the temporal evolution of per-frame features for video-based facial expression recognition. Our reported results demonstrate the advantage of pooling image-set features temporally by stacking the designed manifold network of covariance pooling on top of convolutional network layers.

/pdf/covariance-pooling-for-facial-expression-recognition-3e95h7wy0j.pdf

Covariance Pooling for Facial Expression Recognition

In this paper, we analyze two popular network compression techniques, i.e. filter pruning and low-rank decomposition, in a unified sense. By simply changing the way the sparsity regularization is enforced, filter pruning and low-rank decomposition can be derived accordingly. This provides another flexible choice for network compression because the techniques complement each other. For example, in popular network architectures with shortcut connections (e.g. ResNet), filter pruning cannot deal with the last convolutional layer in a ResBlock while the low-rank decomposition methods can. In addition, we propose to compress the whole network jointly instead of in a layer-wise manner. Our approach proves its potential as it compares favorably to the state-of-the-art on several benchmarks. Code is available at https://github.com/ofsoundof/group_sparsity.

/pdf/group-sparsity-the-hinge-between-filter-pruning-and-3efr0idpqi.pdf

Group Sparsity: The Hinge Between Filter Pruning and Decomposition for Network Compression

As the popularity of mobile photography is growing constantly, lots of efforts are being invested now into building complex hand-crafted camera ISP solutions. In this work, we demonstrate that even the most sophisticated ISP pipelines can be replaced with a single end-to-end deep learning model trained without any prior knowledge about the sensor and optics used in a particular device. For this, we present PyNET, a novel pyramidal CNN architecture designed for fine-grained image restoration that implicitly learns to perform all ISP steps such as image demosaicing, denoising, white balancing, color and contrast correction, demoireing, etc. The model is trained to convert RAW Bayer data obtained directly from mobile camera sensor into photos captured with a professional high-end DSLR camera, making the solution independent of any particular mobile ISP implementation. To validate the proposed approach on the real data, we collected a large-scale dataset consisting of 10 thousand full-resolution RAW-RGB image pairs captured in the wild with the Huawei P20 cameraphone (12.3 MP Sony Exmor IMX380 sensor) and Canon 5D Mark IV DSLR. The experiments demonstrate that the proposed solution can easily get to the level of the embedded P20's ISP pipeline that, unlike our approach, is combining the data from two (RGB + B/W) camera sensors. The dataset, pretrained models and codes used in this paper are available on the project website: https://people.ee.ethz.ch/~ihnatova/pynet.html.

/pdf/replacing-mobile-camera-isp-with-a-single-deep-learning-368ker94ov.pdf

Replacing Mobile Camera ISP with a Single Deep Learning Model

Semantic instance segmentation remains a challenge. We propose to tackle the problem with a discriminative loss function, operating at pixel level, that encourages a convolutional network to produce a representation of the image that can easily be clustered into instances with a simple post-processing step. Our approach of combining an offthe- shelf network with a principled loss function inspired by a metric learning objective is conceptually simple and distinct from recent efforts in instance segmentation and is well-suited for real-time applications. In contrast to previous works, our method does not rely on object proposals or recurrent mechanisms and is particularly well suited for tasks with complex occlusions. A key contribution of our work is to demonstrate that such a simple setup without bells and whistles is effective and can perform on-par with more complex methods. We achieve competitive performance on the Cityscapes segmentation benchmark.

https://openaccess.thecvf.com/content_cvpr_2017_workshops/w5/papers/De_Brabandere_Semantic_Instance_Segmentation_CVPR_2017_paper.pdf

Semantic Instance Segmentation for Autonomous Driving

In this paper we will present the concept of a system that allows for an evolutionary introduction of depth perception into the existing 2D digital TV framework. The work is part of the European Information Society Technologies (IST) project “Advanced Three-Dimensional Television System Technologies” (ATTEST), an activity where industries, research centers and universities have joined forces to design a backwardscompatible, flexible and modular broadcast 3D-TV system, where all parts of the 3D processing chain are optimised to one another. This includes content creation, coding and transmission, display and research in human 3D perception, which be will used to guide the development process. The goals of the project comprise the development of a novel broadcast 3D camera, algorithms to convert existing 2D-video material into 3D, a 2Dcompatible coding and transmission scheme for 3D-video using MPEG2/4/7 technologies and the design of two new autostereoscopic displays.

Luc Van Gool

Papers

Covariance Pooling for Facial Expression Recognition

Group Sparsity: The Hinge Between Filter Pruning and Decomposition for Network Compression

Replacing Mobile Camera ISP with a Single Deep Learning Model

Semantic Instance Segmentation for Autonomous Driving

IST project ATTEST - an evolutionary and optimised approach on 3D-TV