Luc Van Gool

Bio: Luc Van Gool is an academic researcher from Katholieke Universiteit Leuven. The author has contributed to research in topics: Computer science & Object detection. The author has an hindex of 133, co-authored 1307 publications receiving 107743 citations. Previous affiliations of Luc Van Gool include Microsoft & ETH Zurich.

Euclidean 3D Reconstruction from Image Sequences with Variable Focal Lenghts

Abstract: One of the main problems to obtain a Euclidean 3D reconstruction from multiple views is the calibration of the camera. Explicit calibration is not always practical and has to be repeated regularly. Sometimes it is even impossible (i.e. for pictures taken by an unknown camera of an unknown scene). The second possibility is to do auto-calibration. Here the rigidity of the scene is used to obtain constraints on the camera parameters. Existing approaches of this second strand impose that the camera parameters stay exactly the same between different views. This can be very limiting since it excludes changing the focal length to zoom or focus. The paper describes a reconstruction method that allows to vary the focal length. Instead of using one camera one can also use a stereo rig following similar principles, and in which case also reconstruction from a moving rig becomes possible even for pure translation. Synthetic data were used to see how resistant the algorithm is to noise. The results are satisfactory. Also results for a real scene were convincing.

Building Deep Networks on Grassmann Manifolds

Abstract: Learning representations on Grassmann manifolds is popular in quite a few visual recognition tasks. In order to enable deep learning on Grassmann manifolds, this paper proposes a deep network architecture by generalizing the Euclidean network paradigm to Grassmann manifolds. In particular, we design full rank mapping layers to transform input Grassmannian data to more desirable ones, exploit re-orthonormalization layers to normalize the resulting matrices, study projection pooling layers to reduce the model complexity in the Grassmannian context, and devise projection mapping layers to respect Grassmannian geometry and meanwhile achieve Euclidean forms for regular output layers. To train the Grassmann networks, we exploit a stochastic gradient descent setting on manifolds of the connection weights, and study a matrix generalization of backpropagation to update the structured data. The evaluations on three visual recognition tasks show that our Grassmann networks have clear advantages over existing Grassmann learning methods, and achieve results comparable with state-of-the-art approaches.

TriCoS: a tri-level class-discriminative co-segmentation method for image classification

Abstract: The aim of this paper is to leverage foreground segmentation to improve classification performance on weakly annotated datasets – those with no additional annotation other than class labels. We introduce TriCoS, a new co-segmentation algorithm that looks at all training images jointly and automatically segments out the most class-discriminative foregrounds for each image. Ultimately, those foreground segmentations are used to train a classification system. TriCoS solves the co-segmentation problem by minimizing losses at three different levels: the category level for foreground/background consistency across images belonging to the same category, the image level for spatial continuity within each image, and the dataset level for discrimination between classes. In an extensive set of experiments, we evaluate the algorithm on three benchmark datasets: the UCSD-Caltech Birds-200-2010, the Stanford Dogs, and the Oxford Flowers 102. With the help of a modern image classifier, we show superior performance compared to previously published classification methods and other co-segmentation methods.

Real-time sign language letter and word recognition from depth data

Abstract: In this work, we present a system for recognizing letters and finger-spelled words of the American sign language (ASL) in real-time. To this end, the system segments the hand and estimates the hand orientation from captured depth data. The letter classification is based on average neighborhood margin maximization and relies on the segmented depth data of the hands. For word recognition, the letter confidences are aggregated. Furthermore, the word recognition is used to improve the letter recognition by updating the training examples of the letter classifiers on-line.

Color-based object tracking in multi-camera environments

Abstract: This paper presents a multi-view tracker, meant to operate in smart rooms that are equipped with multiple cameras. The cameras are assumed to be calibrated 1 . In particular, we demonstrate a virtual classroom application, where the system automatically selects the camera with the 'best' view on the face of a person moving in the room. Real-time object tracking, which is needed to achieve this, is implemented by means of color-based particle filtering. The use of multiple model histograms for the target (human head) results robust tracking, even when the view on the target changes considerably like from the front to the back. Information is shared between the cameras, which adds robustness to the system. Once one camera has lost the target, it can be reinitialized with the help of the epipolar constraints suggested by the others. Experiments in our research environment corroborate the effectiveness of the approach.

Deep Residual Learning for Image Recognition

Abstract: 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.

Very Deep Convolutional Networks for Large-Scale Image Recognition

Abstract: 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

Abstract: 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.

Deep Residual Learning for Image Recognition

Abstract: 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.

Going deeper with convolutions

Abstract: 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.