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

In this paper, we describe the prototype of an interactive museum guide. It runs on a tablet PC that features a touchscreen, a webcam and a Bluetooth receiver. This guide recognises objects on display in museums based on images of the latter which are taken directly by the visitor. Furthermore, the computer can determine the visitor’s location by receiving signals emitted from Bluetooth senders in the museum, so called BTnodes. This information is used to reduce the search space for the extraction of relevant objects. Hence, the recognition accuracy is increased and the search time reduced. Moreover, this information can be used to indicate the user’s current location in the museum. The prototype has been demonstrated to visitors of the Swiss National Museum in Zurich.

Interactive Museum Guide

In this paper an approach is presented that obtains virtual models from sequences of images. The system can deal with uncalibrated image sequences acquired with a hand held camera. Based on tracked or matched features the relations between multiple views are computed. From this both the structure of the scene and the motion of the camera are retrieved. The ambiguity on the reconstruction is restricted from projective to metric through auto-calibration. A flexible multi-view stereo matching scheme is used to obtain a dense estimation of the surface geometry. From the computed data virtual models can be constructed or, inversely, virtual models can be included in the original images.

/pdf/structure-and-motion-from-image-sequences-4cohhx8mi3.pdf

Structure and motion from image sequences

In this paper, we explore region-based 3D representations of the human face We begin by noting that although they serve as a key ingredient in many state-of-the-art 3D face reconstruction algorithms, very little research has gone into devising strategies for optimally designing them In fact, the great majority of such models encountered in the literature is based on manual segmentations of the face into subregions We propose algorithms that are capable of automatically finding the optimal subdivision given a training set and the number of desired regions The generality of the segmentation approach is demonstrated on examples from the TOSCA database, and a cross-validation experiment on facial data shows that part-based models designed using the proposed algorithms are capable of outperforming alternative segmentations wrt reconstruction accuracy

Optimal regions for linear model-based 3D face reconstruction

https://pure.hw.ac.uk/ws/files/785362/Automation_in_Construction_Aschwanden_IX.pdf

Empiric design evaluation in urban planning

Considerable progress has been made in semantic scene understanding of road scenes with monocular cameras. It is, however, mainly focused on certain specific classes such as cars, bicyclists and pedestrians. This work investigates traffic cones, an object category crucial for traffic control in the context of autonomous vehicles. 3D object detection using images from a monocular camera is intrinsically an ill-posed problem. In this work, we exploit the unique structure of traffic cones and propose a pipelined approach to solve this problem. Specifically, we first detect cones in images by a modified 2D object detector. Following which the keypoints on a traffic cone are recognized with the help of our deep structural regression network, here, the fact that the cross-ratio is projection invariant is leveraged for network regularization. Finally, the 3D position of cones is recovered via the classical Perspective n-Point algorithm using correspondences obtained from the keypoint regression. Extensive experiments show that our approach can accurately detect traffic cones and estimate their position in the 3D world in real time. The proposed method is also deployed on a real-time, autonomous system. It runs efficiently on the low-power Jetson TX2, providing accurate 3D position estimates, allowing a race-car to map and drive autonomously on an unseen track indicated by traffic cones. With the help of robust and accurate perception, our race-car won both Formula Student Competitions held in Italy and Germany in 2018, cruising at a top speed of 54 km/h on our driverless platform “gotthard driverless” https://youtu. be/HegmIXASKow?t==11694. Visualization of the complete pipeline, mapping and navigation can be found on our project page http://people.ee.ethz.ch/=tracczucrich/Trnfflcf.lonc/.

Luc Van Gool

Papers

Interactive Museum Guide

Structure and motion from image sequences

Optimal regions for linear model-based 3D face reconstruction

Empiric design evaluation in urban planning

Real-time 3D Traffic Cone Detection for Autonomous Driving