Abstract: Diversity of food and its attributes represents the culinary habits of peoples from different countries. Thus, this paper addresses the problem of identifying food culture of people around the world and its flavor by classifying two main food attributes, cuisine and flavor. A deep learning model based on multi-scale convotuional networks is proposed for extracting more accurate features from input images. The aggregation of multi-scale convolution layers with different kernel size is also used for weighting the features results from different scales. In addition, a joint loss function based on Negative Log Likelihood (NLL) is used to fit the model probability to multi labeled classes for multi-modal classification task. Furthermore, this work provides a new dataset for food attributes, so-called Yummly48K, extracted from the popular food website, Yummly. Our model is assessed on the constructed Yummly48K dataset. The experimental results show that our proposed method yields 65% and 62% average F1 score on validation and test set which outperforming the state-of-the-art models.

Abstract: Vision-based object detectors are crucial for different applications. They rely on learnt object models. Ideally, we would like to deploy our vision system in the scenario where it must operate, and lead it to self-learn how to distinguish the objects of interest, i.e., without human intervention. However, the learning of each object model requires labelled samples collected through a tiresome manual process. For instance, we are interested in exploring the self-training of a pedestrian detector for driver assistance systems. Our first approach to avoid manual labelling consisted in the use of samples coming from realistic computer graphics, so that their labels are automatically available [12]. This would make possible the desired self-training of our pedestrian detector. However, as we showed in [14], between virtual and real worlds it may be a dataset shift. In order to overcome it, we propose the use of unsupervised domain adaptation techniques that avoid human intervention during the adaptation process. In particular, this paper explores the use of the transductive SVM (T-SVM) learning algorithm in order to adapt virtual and real worlds for pedestrian detection (Fig. 1).

Abstract: This paper proposes a new version of Particle Filter, called Articulated Particle Filter – ArPF -, which has been specifically designed for an efficient sampling of hierarchical spaces, generated by articulated objects. Our approach decomposes the articulated motion into layers for efficiency purposes, making use of a careful modeling of the diffusion noise along with its propagation through the articulations. This produces an increase of accuracy and prevent for divergences. The algorithm is tested on hand tracking due to its complex hierarchical articulated nature. With this purpose, a new dataset generation tool for quantitative evaluation is also presented in this paper.

Abstract: Unconstrained handwritten text recognition systems maximize the combination of two separate probability scores. The first one is the observation probability that indicates how well the returned word sequence matches the input image. The second score is the probability that reflects how likely a word sequence is according to a language model. Current state-of-the-art recognition systems use statistical language models in form of bigram word probabilities. This paper proposes to model the target language by means of a recurrent neural network with long-short term memory cells. Because the network is recurrent, the considered context is not limited to a fixed size especially as the memory cells are designed to deal with long-term dependencies. In a set of experiments conducted on the IAM off-line database we show the superiority of the proposed language model over statistical n-gram models.

Abstract: We present a Bag-of-Visual-and-Depth-Words (BoVDW) model for gesture recognition, an extension of the Bag-of-Visual-Words (BoVW) model, that benefits from the multimodal fusion of visual and depth features. State-of-the-art RGB and depth features, including a new proposed depth descriptor, are analysed and combined in a late fusion fashion. The method is integrated in a continuous gesture recognition pipeline, where Dynamic Time Warping (DTW) algorithm is used to perform prior segmentation of gestures. Results of the method in public data sets, within our gesture recognition pipeline, show better performance in comparison to a standard BoVW model.

Abstract: In this paper, we propose a new approach to extract text regions from graphical documents. In our method, we first empirically construct two sequences of learned dictionaries for the text and graphical parts respectively. Then, we compute the sparse representations of all different sizes and non-overlapped document patches in these learned dictionaries. Based on these representations, each patch can be classified into the text or graphic category by comparing its reconstruction errors. Same-sized patches in one category are then merged together to define the corresponding text or graphic layers which are combined to createfinal text/graphic layer. Finally, in a post-processing step, text regions are further filtered out by using some learned thresholds.

Abstract: Automatic 3D reconstruction of human poses from monocular images is a challenging and popular topic in the computer vision community, which provides a wide range of applications in multiple areas. Solutions for 3D pose estimation involve various learning approaches, such as support vector machines and Gaussian processes, but many encounter difficulties in cluttered scenarios and require additional input data, such as silhouettes, or controlled camera settings. We present a framework that is capable of estimating the 3D pose of a person from single images or monocular image sequences without requiring background information and which is robust to camera variations. The framework models the non-linearity present in human pose estimation as it benefits from flexible learning approaches, including a highly customizable 2D detector. Results on the HumanEva benchmark show how they perform and influence the quality of the 3D pose estimates.

Abstract: World Health Organization estimates that 80% of the world population is affected of back pain during his life. Current practices to analyze back problems are expensive, subjective, and invasive. In this work, we propose a novel tool for posture and range of movement estimation based on the analysis of 3D information from depth maps. Given a set of keypoints defined by the user, RGB and depth data are aligned, depth surface is reconstructed, keypoints are matching using a novel point-to-point fitting procedure, and accurate measurements about posture, spinal curvature, and range of movement are computed. The system shows high precision and reliable measurements, being useful for posture reeducation purposes to prevent musculoskeletal disorders, such as back pain, as well as tracking the posture evolution of patients in rehabilitation treatments.