Abstract: Mode detection in online handwritten documents refers to the process of distinguishing different types of contents, such as text, formulas, diagrams, or tables, one from another. In this paper a new approach to mode detection is proposed that uses bidirectional long-short term memory (BLSTM) neural networks. The BLSTM neural network is a novel type of recursive neural network that has been successfully applied in speech and handwriting recognition. In this paper we show that it has the potential to significantly outperform traditional methods for mode detection, which are usually based on stroke classification. As a further advantage over previous approaches, the proposed system is trainable and does not rely on user-defined heuristics. Moreover, it can be easily adapted to new or additional types of modes by just providing the system with new training data.

Abstract: We present a framework for object segmentation using depth maps based on Random Forest and Graph-cuts theory, and apply it to the segmentation of human limbs. First, from a set of random depth features, Random Forest is used to infer a set of label probabilities for each data sample. This vector of probabilities is used as unary term in α−β swap Graph-cuts algorithm. Moreover, depth values of spatio-temporal neighboring data points are used as boundary potentials. Results on a new multi-label human depth data set show high performance in terms of segmentation overlapping of the novel methodology compared to classical approaches.

Abstract: Head pose estimation is a critical problem in many computer vision applications. These include human computer interaction, video surveillance, face and expression recognition. In most prior work on heads pose estimation, the positions of the faces on which the pose is to be estimated are specified manually. Therefore, the results are reported without studying the effect of misalignment. We propose a method based on partial least squares (PLS) regression to estimate pose and solve the alignment problem simultaneously. The contributions of this paper are two-fold: 1) we show that the kernel version of PLS (kPLS) achieves better than state-of-the-art results on the estimation problem and 2) we develop a technique to reduce misalignment based on the learned PLS factors.

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: Representing a graph with a feature vector is a common way of making statistical machine learning algorithms applicable to the domain of graphs. Such a transition from graphs to vectors is known as graphembedding. A key issue in graphembedding is to select a proper set of features in order to make the vectorial representation of graphs as strong and discriminative as possible. In this article, we propose features that are constructed out of frequencies of node label representatives. We first build a large set of features and then select the most discriminative ones according to different ranking criteria and feature transformation algorithms. On different classification tasks, we experimentally show that only a small significant subset of these features is needed to achieve the same classification rates as competing to state-of-the-art methods.

Abstract: In this article, we focus on the problem of large-scale instance-level image retrieval. For efficiency reasons, it is common to represent an image by a fixed-length descriptor which is subsequently encoded into a small number of bits. We note that most encoding techniques include an unsupervised dimensionality reduction step. Our goal in this work is to learn a better subspace in a supervised manner. We especially raise the following question: “can category-level labels be used to learn such a subspace?” To answer this question, we experiment with four learning techniques: the first one is based on a metric learning framework, the second one on attribute representations, the third one on Canonical Correlation Analysis (CCA) and the fourth one on Joint Subspace and Classifier Learning (JSCL). While the first three approaches have been applied in the past to the image retrieval problem, we believe we are the first to show the usefulness of JSCL in this context. In our experiments, we use ImageNet as a source of category-level labels and report retrieval results on two standard dataseis: INRIA Holidays and the University of Kentucky benchmark. Our experimental study shows that metric learning and attributes do not lead to any significant improvement in retrieval accuracy, as opposed to CCA and JSCL. As an example, we report on Holidays an increase in accuracy from 39.3% to 48.6% with 32-dimensional representations. Overall JSCL is shown to yield the best results.