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

This paper reviews the first challenge on image dehazing (restoration of rich details in hazy image) with focus on proposed solutions and results. The challenge had 2 tracks. Track 1 employed the indoor images (using I-HAZE dataset), while Track 2 outdoor images (using O-HAZE dataset). The hazy images have been captured in presence of real haze, generated by professional haze machines. I-HAZE dataset contains 35 scenes that correspond to indoor domestic environments, with objects with different colors and specularities. O-HAZE contains 45 different outdoor scenes depicting the same visual content recorded in haze-free and hazy conditions, under the same illumination parameters. The dehazing process was learnable through provided pairs of haze-free and hazy train images. Each track had ~ 120 registered participants and 21 teams competed in the final testing phase. They gauge the state-of-the-art in image dehazing.

/pdf/ntire-2018-challenge-on-image-dehazing-methods-and-results-2auhtl0o6b.pdf

NTIRE 2018 Challenge on Image Dehazing: Methods and Results

In this paper, we describe the application of the novel SURF (Speeded Up Robust Features) algorithm [1] for the recognition of objects of art. For this purpose, we developed a prototype of a mobile interactive museum guide consisting of a tablet PC that features a touchscreen and a webcam. This guide recognises objects in museums based on images taken by the visitor. Using different image sets of real museum objects, we demonstrate that both the object recognition performance as well as the speed of the SURF algorithm surpasses the results obtained with SIFT, its main contender.

Interactive museum guide : fast and robust recognition of museum objects

This work addresses the problem of semantic scene understanding under dense fog. Although considerable progress has been made in semantic scene understanding, it is mainly related to clear-weather scenes. Extending recognition methods to adverse weather conditions such as fog is crucial for outdoor applications. In this paper, we propose a novel method, named Curriculum Model Adaptation (CMAda), which gradually adapts a semantic segmentation model from light synthetic fog to dense real fog in multiple steps, using both synthetic and real foggy data. In addition, we present three other main stand-alone contributions: (1) a novel method to add synthetic fog to real, clear-weather scenes using semantic input; (2) a new fog density estimator; (3) the Foggy Zurich dataset comprising 3808 real foggy images, with pixel-level semantic annotations for 16 images with dense fog. Our experiments show that (1) our fog simulation slightly outperforms a state-of-the-art competing simulation with respect to the task of semantic foggy scene understanding (SFSU); (2) CMAda improves the performance of state-of-the-art models for SFSU significantly by leveraging unlabeled real foggy data. The datasets and code will be made publicly available.

/pdf/model-adaptation-with-synthetic-and-real-data-for-semantic-3uy4iuzpgi.pdf

Model Adaptation with Synthetic and Real Data for Semantic Dense Foggy Scene Understanding

We propose a method for object detection in cluttered real images, given a single hand-drawn example as model. The image edges are partitioned into contour segments and organized in an image representation which encodes their interconnections: the Contour Segment Network. The object detection problem is formulated as finding paths through the network resembling the model outlines, and a computationally efficient detection technique is presented. An extensive experimental evaluation on detecting five diverse object classes over hundreds of images demonstrates that our method works in very cluttered images, allows for scale changes and considerable intra-class shape variation, is robust to interrupted contours, and is computationally efficient.

https://lirias.kuleuven.be/bitstream/123456789/73490/1/Ferrari_Tuytelaars_VanGool-objcat_csn-eccv06.pdf

Object Detection by Contour Segment Networks

State-of-the-art methods for 3D hand pose estimation from depth images require large amounts of annotated training data. We propose modelling the statistical relationship of 3D hand poses and corresponding depth images using two deep generative models with a shared latent space. By design, our architecture allows for learning from unlabeled image data in a semi-supervised manner. Assuming a one-to-one mapping between a pose and a depth map, any given point in the shared latent space can be projected into both a hand pose or into a corresponding depth map. Regressing the hand pose can then be done by learning a discriminator to estimate the posterior of the latent pose given some depth map. To prevent over-fitting and to better exploit unlabeled depth maps, the generator and discriminator are trained jointly. At each iteration, the generator is updated with the back-propagated gradient from the discriminator to synthesize realistic depth maps of the articulated hand, while the discriminator benefits from an augmented training set of synthesized samples and unlabeled depth maps. The proposed discriminator network architecture is highly efficient and runs at 90fps on the CPU with accuracies comparable or better than state-of-art on 3 publicly available benchmarks.

/pdf/crossing-nets-combining-gans-and-vaes-with-a-shared-latent-53ottz402c.pdf

Luc Van Gool

Papers

NTIRE 2018 Challenge on Image Dehazing: Methods and Results

Interactive museum guide : fast and robust recognition of museum objects

Model Adaptation with Synthetic and Real Data for Semantic Dense Foggy Scene Understanding

Object Detection by Contour Segment Networks

Crossing Nets: Combining GANs and VAEs with a Shared Latent Space for Hand Pose Estimation