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

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

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.

Exemplary systems and methods segment a foreground from a background image in a video sequence. In one implementation, a system refines a segmentation boundary between the foreground and the background image by attenuating background contrast while preserving contrast of the segmentation boundary itself, providing an accurate background cut of live video in real time. A substitute background may then be merged with the segmented foreground within the live video. The system can apply an adaptive background color mixture model to improve segmentation of foreground from background under various background changes, such as camera movement, illumination change, and movement of small objects in the background.

Background removal in a live video

Techniques are disclosed to improve quality of images that may be blurred or underexposed (e.g., because of camera shake, taken in dim lighting conditions, or taken of high action scenes). The techniques may be implemented in a digital camera, digital video camera, or a digital camera capable of capturing video. In one described implementation, a digital camera includes an image sensor, a storage device, and a processing unit. The image sensor captures two images from a same scene which are stored on the storage device. The processing unit enhances the captured images with luminance correction.

Digital cameras with luminance correction

In this paper, we present a novel approach to keyframe-based tracking, called bi-directional tracking. Given two object templates in the beginning and ending keyframes, the bi-directional tracker outputs the MAP (maximum a posterior) solution of the whole state sequence of the target object in the Bayesian framework. First, a number of 3D trajectory segments of the object are extracted from the input video, using a novel trajectory segment analysis. Second, these disconnected trajectory segments due to occlusion are linked by a number of inferred occlusion segments. Last, the MAP solution is obtained by trajectory optimization in a coarse-to-fine manner. Experimental results show the robustness of our approach with respect to sudden motion, ambiguity, and short and long periods of occlusion.

/pdf/bidirectional-tracking-using-trajectory-segment-analysis-as6p1f7zqa.pdf

Bidirectional tracking using trajectory segment analysis

Inverted indexing is a popular non-exhaustive solution to large scale search. An inverted file is built by a quantizer such as k-means or a tree structure. It has been found that multiple inverted files, obtained by multiple independent random quantizers, are able to achieve practically good recall and speed. Instead of computing the multiple quantizers independently, we present a method that creates them jointly. Our method jointly optimizes all code words in all quantizers. Then it assigns these code words to the quantizers. In experiments this method shows significant improvement over various existing methods that use multiple independent quantizers. On the one-billion set of SIFT vectors, our method is faster and more accurate than a recent state-of-the-art inverted indexing method.

/pdf/joint-inverted-indexing-3s0rma6rt6.pdf

Joint Inverted Indexing

A user interface enables interactive image matting to be performed on an image The user interface may provide results including an alpha matte as feedback in real time. The user interface may provide interactive tools for selecting a portion of the image, and an unknown region for alpha matte processing may be automatically generated adjacent to the selected region. The user may interactively refine the alpha matte as desired to obtain a satisfactory result.

Jian Sun

Papers

Background removal in a live video

Digital cameras with luminance correction

Bidirectional tracking using trajectory segment analysis

Joint Inverted Indexing

Interactive image matting