Publications

In this paper we propose a novel image descriptor built by computing the covariance of pixel level features on densely sampled patches and encoding them using their covariance. Appropriate projections to the Euclidean space and feature normalizations are employed in order to provide a strong descriptor usable with linear classifiers. In order to remove border effects, we further enhance the Spatial Pyramid representation with bilinear interpolation. Experimental results conducted on two common datasets for object and texture classification show that the performance of our method is comparable with state of the art techniques, but removing any dataset specific dependency in the feature encoding step.

The characterization of the interacting behaviors of complex biological systems is a primary objective in protein–protein network analysis and computational biology. In this paper we present FunMod, an innovative Cytoscape version 2.8 plugin that is able to mine undirected protein–protein networks and to infer sub-networks of interacting proteins intimately correlated with relevant biological pathways. This plugin may enable the discovery of new pathways involved in diseases. In order to describe the role of each protein within the relevant biological pathways, FunMod computes and scores three topological features of the identified sub-networks. By integrating the results from biological pathway clustering and topological network analysis, FunMod proved to be useful for the data interpretation and the generation of new hypotheses in two case studies.

The categorization of different staining patterns in HEp-2 cell slides by means of indirect immunofluorescence (IIF) is important for the differential diagnosis of autoimmune diseases. The clinical practice usually relies on the visual evaluation of the slides, which is time-consuming and subject to the specialist's experience. Thus, there is a growing demand for computer-aided solutions capable of automatically classifying HEp-2 staining patterns. In the attempt to identify the most suited strategy for this task, in this work we compare two approaches based on Support Vector Machines and Subclass Discriminant Analysis. These techniques classify the available samples, characterized through a limited set of optimal textural attributes that are identified with a feature selection scheme. Our experimental results show that both strategies have a good concordance with the diagnosis of the human specialist and show the better performance of the Subclass Discriminant Analysis (91% accuracy) compared to Support Vector Machines (87% accuracy).

We present a novel method for monocular hand gesture recognition in ego-vision scenarios that deals with static and dynamic gestures and can achieve high accuracy results using a few positive samples. Specifically, we use and extend the dense trajectories approach that has been successfully introduced for action recognition. Dense features are extracted around regions selected by a new hand segmentation technique that integrates superpixel classification, temporal and spatial coherence. We extensively testour gesture recognition and segmentation algorithms on public datasets and propose a new dataset shot with a wearable camera. In addition, we demonstrate that our solution can work in near real-time on a wearable device.

Smith Waterman algorithm (S-W) is nowadays considered one of the best method to perform local alignments of biological sequences characterizing proteins, DNA and RNA molecules. Indeed, S-W is able to ensure better accuracy levels with respect to the heuristic alignment algorithms by extensively exploring all the possible alignment configurations between the sequences under examination. It has been proven that the first amino acid (AA) or nucleotide (NT) inserted/deleted (that identify a gap open) found during the alignment operations performed on sequences is more significant from a biological point of view than the subsequent ones (called gap extension), making the so called Affine Gap model a viable solution for biomolecules alignment. However, this version of S-W algorithm is expensive both in terms of computation as well as in terms of memory requirements with respect to others less demanding solutions such as the ones using a Linear Gap model. In order to overcome these drawbacks we have developed an optimised version of the S-Walgorithm based on Affine Gap model called Dynamic Gap Selector (DGS S-W). Differently from the standard S-W Affine Gap method, the proposed DGS S-W method reduces the memory requirements from 3*N*M to N*M where N and M represents the size of the compared sequences. In terms of computational costs, the proposed algorithm reduces by a factor of 2 the number of operations required by the standard Affine Gap model. DGS S-W method has been tested on two protein and one RNA sequences datasets, showing mapping scores very similar to those reached thanks to the classical S-W Affine Gap method and, at the same time, reduced computational costs and memory usage.

In this paper we address the task of learning how to segment a particular class of objects, by means of a training set of images and their segmentations. In particular we propose a method to overcome the extremely high training time of a previously proposed solution to this problem, Kernelized Structural Support Vector Machines. We employ a one-class SVM working with joint kernels to robustly learn significant support vectors (representative image-mask pairs) and accordingly weight them to build a suitable energy function for the graph cut framework. We report results obtained on two public datasets and a comparison of training times on different training set sizes.