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

The explosion of image data on the Internet has the potential to foster more sophisticated and robust models and algorithms to index, retrieve, organize and interact with images and multimedia data. But exactly how such data can be harnessed and organized remains a critical problem. We introduce here a new database called “ImageNet”, a large-scale ontology of images built upon the backbone of the WordNet structure. ImageNet aims to populate the majority of the 80,000 synsets of WordNet with an average of 500-1000 clean and full resolution images. This will result in tens of millions of annotated images organized by the semantic hierarchy of WordNet. This paper offers a detailed analysis of ImageNet in its current state: 12 subtrees with 5247 synsets and 3.2 million images in total. We show that ImageNet is much larger in scale and diversity and much more accurate than the current image datasets. Constructing such a large-scale database is a challenging task. We describe the data collection scheme with Amazon Mechanical Turk. Lastly, we illustrate the usefulness of ImageNet through three simple applications in object recognition, image classification and automatic object clustering. We hope that the scale, accuracy, diversity and hierarchical structure of ImageNet can offer unparalleled opportunities to researchers in the computer vision community and beyond.

/pdf/imagenet-a-large-scale-hierarchical-image-database-1e8f663s1h.pdf

ImageNet: A large-scale hierarchical image database

There is large consent that successful training of deep networks requires many thousand annotated training samples. In this paper, we present a network and training strategy that relies on the strong use of data augmentation to use the available annotated samples more efficiently. The architecture consists of a contracting path to capture context and a symmetric expanding path that enables precise localization. We show that such a network can be trained end-to-end from very few images and outperforms the prior best method (a sliding-window convolutional network) on the ISBI challenge for segmentation of neuronal structures in electron microscopic stacks. Using the same network trained on transmitted light microscopy images (phase contrast and DIC) we won the ISBI cell tracking challenge 2015 in these categories by a large margin. Moreover, the network is fast. Segmentation of a 512x512 image takes less than a second on a recent GPU. The full implementation (based on Caffe) and the trained networks are available at http://lmb.informatik.uni-freiburg.de/people/ronneber/u-net .

/pdf/u-net-convolutional-networks-for-biomedical-image-2xoyjzggxv.pdf

U-Net: Convolutional Networks for Biomedical Image Segmentation

We describe a method to completely automatically recover 3D scene structure together with 3D camera positions from a sequence of images acquired by an unknown camera undergoing unknown movement. Unlike "tuned" systems which use calibration objects or markers to recover this information, and are therefore often limited to a particular scale, the approach of this paper is more general and can be applied to a large class of scenes. It is demonstrated here for interior and exterior sequences using both controlled-motion and handheld cameras. The paper reviews Computer Vision research into structure and motion recovery, providing a tutorial introduction to the geometry of multiple views, estimation and correspondence in video streams. The core method, which simultaneously extracts the 3D scene structure and camera positions, is applied to the automated recovery of VRML 3D textured models from a video sequence.

/pdf/automatic-3d-model-acquisition-and-generation-of-new-images-44s12wteol.pdf

Automatic 3D model acquisition and generation of new images from video sequences

The goal of this paper is speaker diarisation of videos collected 'in the wild'. We make three key contributions. First, we propose an automatic audio-visual diarisation method for YouTube videos. Our method consists of active speaker detection using audio-visual methods and speaker verification using self-enrolled speaker models. Second, we integrate our method into a semi-automatic dataset creation pipeline which significantly reduces the number of hours required to annotate videos with diarisation labels. Finally, we use this pipeline to create a large-scale diarisation dataset called VoxConverse, collected from 'in the wild' videos, which we will release publicly to the research community. Our dataset consists of overlapping speech, a large and diverse speaker pool, and challenging background conditions.

Spot the Conversation: Speaker Diarisation in the Wild.

In this paper, we describe a simple approach to learning models of visual object categories from images gathered from Internet image search engines. The images for a given keyword are typically highly variable, with a large fraction being unrelated to the query term, and thus pose a challenging environment from which to learn. By training our models directly from Internet images, we remove the need to laboriously compile training data sets, required by most other recognition approaches-this opens up the possibility of learning object category models “on-the-fly.” We describe two simple approaches, derived from the probabilistic latent semantic analysis (pLSA) technique for text document analysis, that can be used to automatically learn object models from these data. We show two applications of the learned model: first, to rerank the images returned by the search engine, thus improving the quality of the search engine; and second, to recognize objects in other image data sets.

/pdf/learning-object-categories-from-internet-image-searches-3y509ckf0t.pdf

Learning Object Categories From Internet Image Searches

We address the task of accurately localizing the eyes in face images extracted by a face detector, an important problem to be solved because of the negative effect of poor localization on face recognition accuracy. We investigate three approaches to the task: a regression approach aiming to directly minimize errors in the predicted eye positions, a simple Bayesian model of eye and non-eye appearance, and a discriminative eye detector trained using AdaBoost. By using identical training and test data for each method we are able to perform an unbiased comparison. We show that, perhaps surprisingly, the simple Bayesian approach performs best on databases including challenging images, and performance is comparable to more complex state-of-the-art methods.

Regression and Classification Approaches to Eye Localization in Face Images.

There has been a quantum leap in the performance of automated lip reading recently due to the application of neural network sequence models trained on a very large corpus of aligned text and face videos. However, this advance has only been demonstrated for frontal or near frontal faces, and so the question remains: can lips be read in profile to the same standard? The objective of this paper is to answer that question. We make three contributions: first, we obtain a new large aligned training corpus that contains profile faces, and select these using a face pose regressor network; second, we propose a curriculum learning procedure that is able to extend SyncNet [10] (a network to synchronize face movements and speech) progressively from frontal to profile faces; third, we demonstrate lip reading in profile for unseen videos. The trained model is evaluated on a held out test set, and is also shown to far surpass the state of the art on the OuluVS2 multi-view benchmark.

/pdf/lip-reading-in-profile-5froekdvqb.pdf

Andrew Zisserman

Papers

Automatic 3D model acquisition and generation of new images from video sequences

Spot the Conversation: Speaker Diarisation in the Wild.

Learning Object Categories From Internet Image Searches

Regression and Classification Approaches to Eye Localization in Face Images.

Lip Reading in Profile.