Eye gaze detection under challenging lighting conditions is a non-trivial task. Pixel intensity and the shades around the eye region may change depending on the time of day, location, or due to artificial lighting. This paper introduces a lighting-adaptive solution for robust eye gaze detection. First, we propose a binarization and cropping technique to limit our region of interest. Then we dev...

Several algorithms have been proposed for discovering patterns from trajectories of moving objects, but only a few have concentrated on outlier detection. Existing approaches, in general, discover spatial outliers, and do not provide any further analysis of the patterns. In this paper we introduce semantic spatial and spatio-temporal outliers and propose a new algorithm for trajectory outlier d...

Detection of objects is in general a computationally demanding task. To simplify the problem it is of interest to focus the attention to a set of regions of interest. Indoor environments often have large homogeneous textured objects, such as walls and furniture. In this paper we present a model which detects large homogeneous regions and uses this information to search for targets that are smal...

We present: {1) A multi-scale representation of grey-level shape, called scale-space primal sketch, which makes explicit both fea­ tu res in scale-s pace and the relations between f eatures at differ­ en t levels of scale. (2} A theory for extraction of significant im­ age structure from this representation. (3} Applications to edge detection, histogram analysis and junction classification demo...

This paper investigates the robustness of a new thermal-infrared pedestrian detection system under different outdoor environmental conditions. In first place the algorithm for pedestrian ROI extraction in thermal-infrared video based on both thermal and motion information is introduced. Then, the evaluation of the proposal is detailed after describing the complete thermal and motion information...

This paper discusses the use of neural networks to locate regions of interest (ROIs) for fingerprint classification using feature-encoded fingerprint images. The target areas are those useful for the classification of fingerprints: whorls, loops, arches, and deltas. Our approach is to limit the amount of data which a classification algorithm must consider by determining with high accuracy those...

A novel technique to detect regions of interest in a time series as deviation from the characteristic behavior is proposed. The deterministic form of a signal is obtained using a reliably trained MLP neural network with detailed complexity management and cross-validation based generalization assurance. The proposed technique is demonstrated with simulated and real data.

Driven by increasing concerns over automotive safety, the development of on-board automotive driver assistance systems to alert drivers about potential collisions with objects such as pedestrians or vehicles has become a huge area of research. The aim of this paper is to present a novel approach for the first processing stage in many of these types of systems. An algorithm is presented which us...