Falls are one of the major risks of injury for elderly living alone at home. Computer vision-based systems offer a new, low-cost and promising solution for fall detection. This paper presents a new fall-detection tool, based on a commercial RGB-D camera. The proposed system is capable of accurately detecting several types of falls, performing a real time algorithm in order to determine whether a fall has occurred. The proposed approach is based on evaluating the contraction and the expansion speed of the width, height and depth of the 3D human bounding box, as well as its position in the space. Our solution requires no pre-knowledge of the scene (i.e. the recognition of the floor in the virtual environment) with the only constraint about the knowledge of the RGB-D camera position in the room. Moreover, the proposed approach is able to avoid false positive as: sitting, lying down, retrieve something from the floor. Experimental results qualitatively and quantitatively show the quality of the proposed approach in terms of both robustness and background and speed independence.

We propose in this paper a modification of one of the modern state-of-the-art genetic programming algorithms used for data-driven modeling, namely the Bi-objective Genetic Programming (BioGP). The original method is based on a concurrent minimization of both the training error and complexity of multiple candidate models encoded as Genetic Programming trees. Also, BioGP is empowered by a predator-prey co-evolutionary model where virtual predators are used to suppress solutions (preys) characterized by a poor trade-off error vs complexity. In this work, we incorporate in the original BioGP an adaptive mechanism that automatically tunes the mutation rate, based on a characterization of the current population (in terms of entropy) and on the information that can be extracted from it. We show through numerical experiments on two different datasets from the energy domain that the proposed method, named BioAGP (where "A" stands for "Adaptive"), performs better than the original BioGP, allowing the search to maintain a good diversity level in the population, without affecting the convergence rate.