Arnaldo J. Abrantes

A class of constrained clustering algorithms for object boundary extraction

Boundary extraction is a key task in many image analysis operations. This paper describes a class of constrained clustering algorithms for object boundary extraction that includes several well-known algorithms proposed in different fields (deformable models, constrained clustering, data ordering, and traveling salesman problems). The algorithms belonging to this class are obtained by the minimization of a cost function with two terms: a quadratic regularization term and an image-dependent term defined by a set of weighting functions. The minimization of the cost function is achieved by lowpass filtering the previous model shape and by attracting the model units toward the centroids of their attraction regions. To define a new algorithm belonging to this class, the user has to specify a regularization matrix and a set of weighting functions that control the attraction of the model units toward the data. The usefulness of this approach is twofold: it provides a unified framework for many existing algorithms in pattern recognition and deformable models, and allows the design of new recursive schemes.

Tracking Groups of Pedestrians in Video Sequences

This paper describes an algorithm for tracking groups of objects in video sequences. The main difficulties addressed in this work concern total occlusions of the objects to be tracked as well as group merging and splitting. A two layer solution is proposed to overcome these difficulties. The first layer produces a set of spatio temporal strokes based on low level operations which manage to track the active regions most of the time. The second layer performs a consistent labeling of the detected segments using a statistical model based on Bayesian networks. The Bayesian network is recursively computed during the tracking operation and allows the update of the tracker results everytime new information is available. Experimental tests are included to show the performance of the algorithm in ambiguous situations.

Shape alignment—optimal initial point and pose estimation

Abstract This paper describes an algorithm for the optimal alignment of a pair of 2D shapes defined by point sequences. The proposed algorithm provides closed-form expressions for the estimation of the initial point, scale and pose parameters.

Long term tracking using Bayesian networks

This paper addresses long term tracking of multiple objects with occlusions. Bayesian networks are used to model the interaction among the detected tracks and for conflict management, allowing the tracker to update the labelling decisions as soon as new information is available. If several objects overlap in the image domain and then become separated in the next frames, the proposed algorithm is able to accumulate the evidence extracted from the images and to disambiguate the competing labels. The system also provides a confidence degree for each labelling decision. Experimental results are provided to illustrate the performance of the proposed method for long term tracking of multiple pedestrians.

A Method for Dynamic Clustering of Data

This paper describes a method for the segmentation of dynamic data. It extends well known algorithms developed in the context of static clustering (e.g., the c-means algorithm, Kohonen maps, elastic nets and fuzzy c-means). The work is based on an unified framework for constrained clustering recently proposed by the authors in [1]. This framework is extended by using a motion model for the clusters which includes global and local evolution of the data centroids. A noise model is also proposed to increase the robustness of the dynamic clustering algorithm with respect to outliers.

Hybrid harmonic coding of speech at low bit-rates

Abstract This paper presents a novel approach to sinusoidal coding of speech which avoids the use of a voicing detector. The proposed model represents the speech signal as a sum of sinusoids and bandpass random signals and it is denoted hybrid harmonic model in this paper. The use of two different sets of basis functions increases the robustness of the model since there is no need to switch between techniques tailored to particular classes of sounds. Sinusoidal basis functions with harmonically related frequencies allow an accurate representation of the quasi-periodic structure of voiced speech but show difficulties in representing unvoiced sounds. On the other hand, the bandpass random functions are well suited for high quality representation of unvoiced speech sounds, since their bandwidth is larger than the bandwidth of sinusoids. The amplitudes of both sets of basis functions are simultaneously estimated by a least squares algorithm and the output speech signal is synthesized in the time domain by the superposition of all basis functions multiplied by their amplitudes. Experimental tests confirm an improved performance of the hybrid model for operation with noise-corrupted input speech, relative to classic sinusoidal models which exhibit a strong dependency on voicing decision. Finally, the implementation and test of a fully quantized hybrid coder at 4.8 kbit/s is described.

Methods for automatic and assisted image annotation

Personal memories composed of digital pictures are very popular at the moment. To retrieve these media items annotation is required. During the last years, several approaches have been proposed in order to overcome the image annotation problem. This paper presents our proposals to address this problem. Automatic and semi-automatic learning methods for semantic concepts are presented. The automatic method is based on semantic concepts estimated using visual content, context metadata and audio information. The semi-automatic method is based on results provided by a computer game. The paper describes our proposals and presents their evaluations.

Playing games as a way to improve automatic image annotation

Image annotation is hard to do in an automatic way. In this paper, we propose a framework for image annotation that combines the benefits of three paradigms: automatic annotation, human intervention and entertainment activities. We also describe our proposal inside this framework, the ASAA (application for semi-automatic annotation) interface, a new computer game for image tagging. The application has a 3D game interface, and is supported by a game engine that uses a system for automatic image classification and gestural input to play the game. We present results of the performance of semantic models obtained with a training set enlarged by images annotated during the game activity as well as usability tests of the application.

Automatic Vehicle Detection and Classification

This paper presents a proposal for an automatic vehicle detection and classification (AVDC) system. The proposed AVDC should classify vehicles accordingly to the Portuguese legislation (vehicle height over the first axel and number of axels), and should also support profile based classification. The AVDC should also fulfill the needs of the Portuguese motorway operator, Brisa. For the classification based on the profile we propose the use of Eigenprofiles, a technique based on Principal Components Analysis. The system should also support multi-lane free flow for future integration in this kind of environments.

Estimation of the Bayesian network architecture for object tracking in video sequences

It was recently proposed the use of Bayesian networks for object tracking. Bayesian networks allow modeling the interaction among detected trajectories, in order to obtain reliable object identification in the presence of occlusions. However, the architecture of the Bayesian network has been defined using simple heuristic rules, which fail in many cases. This paper addresses the above problem and presents a new method to estimate the network architecture from the video sequences using supervised learning techniques. Experimental results are presented showing that significant performance gains (increase of accuracy and decrease of complexity) are achieved by the proposed methods.