Abstract: Before the big data era, scene recognition was often approached with two-step inference using localized intermediate representations (objects, topics, and so on). One of such approaches is the semantic manifold (SM), in which patches and images are modeled as points in a semantic probability simplex. Patch models are learned resorting to weak supervision via image labels, which leads to the problem of scene categories co-occurring in this semantic space. Fortunately, each category has its own co-occurrence patterns that are consistent across the images in that category. Thus, discovering and modeling these patterns are critical to improve the recognition performance in this representation. Since the emergence of large data sets, such as ImageNet and Places, these approaches have been relegated in favor of the much more powerful convolutional neural networks (CNNs), which can automatically learn multi-layered representations from the data. In this paper, we address many limitations of the original SM approach and related works. We propose discriminative patch representations using neural networks and further propose a hybrid architecture in which the semantic manifold is built on top of multiscale CNNs. Both representations can be computed significantly faster than the Gaussian mixture models of the original SM. To combine multiple scales, spatial relations, and multiple features, we formulate rich context models using Markov random fields. To solve the optimization problem, we analyze global and local approaches, where a top-down hierarchical algorithm has the best performance. Experimental results show that exploiting different types of contextual relations jointly consistently improves the recognition accuracy.

Abstract: Kidney stone formation is a common disease and the incidence rate is constantly increasing worldwide. It has been shown that the classification of kidney stones can lead to an important reduction of the recurrence rate. The classification of kidney stones by human experts on the basis of certain visual color and texture features is one of the most employed techniques. However, the knowledge of how to analyze kidney stones is not widespread, and the experts learn only after being trained on a large number of samples of the different classes. In this paper we describe a new device specifically designed for capturing images of expelled kidney stones, and a method to learn and apply the experts knowledge with regard to their classification. We show that with off the shelf components, a carefully selected set of features and a state of the art classifier it is possible to automate this difficult task to a good degree. We report results on a collection of 454 kidney stones, achieving an overall accuracy of 63% for a set of eight classes covering almost all of the kidney stones taxonomy. Moreover, for more than 80% of samples the real class is the first or the second most probable class according to the system, being then the patient recommendations for the two top classes similar. This is the first attempt towards the automatic visual classification of kidney stones, and based on the current results we foresee better accuracies with the increase of the dataset size.

Abstract: State-of-the-art approaches on hand pose estimation from depth images have reported promising results under quite controlled considerations. In this paper we propose a two-step pipeline for recovering the hand pose from a sequence of depth images. The pipeline has been designed to deal with images taken from any viewpoint and exhibiting a high degree of finger occlusion. In a first step we initialize the hand pose using a part-based model, fitting a set of hand components in the depth images. In a second step we consider temporal data and estimate the parameters of a trained bilinear model consisting of shape and trajectory bases. Results on a synthetic, highly-occluded dataset demonstrate that the proposed method outperforms most recent pose recovering approaches, including those based on CNNs.

Abstract: Despite recent significant advances, object detection continues to be an extremely challenging problem in real scenarios. In order to develop a detector that successfully operates under these conditions, it becomes critical to leverage upon multiple cues, multiple imaging modalities, and a strong multiview (MV) classifier that accounts for different object views and poses. In this paper, we provide an extensive evaluation that gives insight into how each of these aspects (multicue, multimodality, and strong MV classifier) affect accuracy both individually and when integrated together. In the multimodality component, we explore the fusion of RGB and depth maps obtained by high-definition light detection and ranging, a type of modality that is starting to receive increasing attention. As our analysis reveals, although all the aforementioned aspects significantly help in improving the accuracy, the fusion of visible spectrum and depth information allows to boost the accuracy by a much larger margin. The resulting detector not only ranks among the top best performers in the challenging KITTI benchmark, but it is built upon very simple blocks that are easy to implement and computationally efficient.

Abstract: Optical Music Recognition is the task of transcribing a music score into a machine readable format. Many music scores are written in a single staff, and therefore, they could be treated as a sequence. Therefore, this work explores the use of Long Short-Term Memory (LSTM) Recurrent Neural Networks for reading the music score sequentially, where the LSTM helps in keeping the context. For training, we have used a synthetic dataset of more than 40000 images, labeled at primitive level

Abstract: A main weakness of virtual bronchoscopic navigation (VBN) is unsuccessful segmentation of distal branches approaching peripheral pulmonary nodules (PPN). CT scan acquisition protocol is pivotal for segmentation covering the utmost periphery. We hypothesize that application of continuous positive airway pressure (CPAP) during CT acquisition could improve visualization and segmentation of peripheral bronchi. The purpose of the present pilot study is to compare quality of segmentations under 4 CT acquisition modes: inspiration (INSP), expiration (EXP) and both with CPAP (INSP-CPAP and EXP-CPAP).

Abstract: Deep learning for human action recognition in videos is making significant progress, but is slowed down by its dependency on expensive manual labeling of large video collections. In this work, we investigate the generation of synthetic training data for action recognition, as it has recently shown promising results for a variety of other computer vision tasks. We propose an interpretable parametric generative model of human action videos that relies on procedural generation and other computer graphics techniques of modern game engines. We generate a diverse, realistic, and physically plausible dataset of human action videos, called PHAV for ”Procedural Human Action Videos”. It contains a total of 39, 982 videos, with more than 1, 000 examples for each action of 35 categories. Our approach is not limited to existing motion capture sequences, and we procedurally define 14 synthetic actions. We introduce a deep multi-task representation learning architecture to mix synthetic and real videos, even if the action categories differ. Our experiments on the UCF101 and HMDB51 benchmarks suggest that combining our large set of synthetic videos with small real-world datasets can boost recognition performance, significantly
outperforming fine-tuning state-of-the-art unsupervised generative models of videos.