Jaime A. Martins

Luminance, Colour, Viewpoint and Border Enhanced Disparity Energy Model

The visual cortex is able to extract disparity information through the use of binocular cells. This process is reflected by the Disparity Energy Model, which describes the role and functioning of simple and complex binocular neuron populations, and how they are able to extract disparity. This model uses explicit cell parameters to mathematically determine preferred cell disparities, like spatial frequencies, orientations, binocular phases and receptive field positions. However, the brain cannot access such explicit cell parameters; it must rely on cell responses. In this article, we implemented a trained binocular neuronal population, which encodes disparity information implicitly. This allows the population to learn how to decode disparities, in a similar way to how our visual system could have developed this ability during evolution. At the same time, responses of monocular simple and complex cells can also encode line and edge information, which is useful for refining disparities at object borders. The brain should then be able, starting from a low-level disparity draft, to integrate all information, including colour and viewpoint perspective, in order to propagate better estimates to higher cortical areas.

Biological models for active vision: towards a unified architecture

Building a general-purpose, real-time active vision system completely based on biological models is a great challenge. We apply a number of biologically plausible algorithms which address different aspects of vision, such as edge and keypoint detection, feature extraction, optical flow and disparity, shape detection, object recognition and scene modelling into a complete system. We present some of the experiments from our ongoing work, where our system leverages a combination of algorithms to solve complex tasks.

Disparity energy model using a trained neuronal population

Depth information using the biological Disparity Energy Model can be obtained by using a population of complex cells. This model explicitly involves cell parameters like their spatial frequency, orientation, binocular phase and position difference. However, this is a mathematical model. Our brain does not have access to such parameters, it can only exploit responses. Therefore, we use a new model for encoding disparity information implicitly by employing a trained binocular neuronal population. This model allows to decode disparity information in a way similar to how our visual system could have developed this ability, during evolution, in order to accurately estimate disparity of entire scenes.

Low-cost Natural Interface Based on Head Movements

Abstract Sometimes people look for freedom in the virtual world. However, not all have the possibility to interact with a computer in the same way. Nowadays, almost every job requires interaction with computerized systems, so people with physical impairments do not have the same freedom to control a mouse, a keyboard or a touchscreen. In the last years, some of the government programs to help people with reduced mobility suffered a lot with the global economic crisis and some of those programs were even cut down to reduce costs. This paper focuses on the development of a touchless human-computer interface, which allows anyone to control a computer without using a keyboard, mouse or touchscreen. By reusing Microsoft Kinect sensors from old videogames consoles, a cost-reduced, easy to use, and open-source interface was developed, allowing control of a computer using only the head, eyes or mouth movements, with the possibility of complementary sound commands. There are already available similar commercial solutions, but they are so expensive that their price tends to be a real obstacle in their purchase; on the other hand, free solutions usually do not offer the freedom that people with reduced mobility need. The present solution tries to address these drawbacks.

Cortical multiscale line-edge disparity model

Most biological approaches to disparity extraction rely on the disparity energy model (DEM). In this paper we present an alternative approach which can complement the DEM model. This approach is based on the multiscale coding of lines and edges, because surface structures are composed of lines and edges and contours of objects often cause edges against their background. We show that the line/edge approach can be used to create a 3D wireframe representation of a scene and the objects therein. It can also significantly improve the accuracy of the DEM model, such that our biological models can compete with some state-of-the-art algorithms from computer vision.

GACN: Self-Clustering Genetic Algorithm for Constrained Networks

Extending the lifespan of a wireless sensor network is a complex problem that involves several factors, ranging from device hardware capacity (batteries, processing capabilities, and radio efficiency) to the chosen software stack, which is often unaccounted for by the previous approaches. This letter proposes a genetic algorithm-based clustering optimization method for constrained networks that significantly improves the previous state-of-the-art results, while accounting for the specificities of the Internet engineering task force, Constrained RESTful Environment (CoRE), standards for data transmission and specifically relying on CoRE interfaces, which fit this purpose very well.

Proto-object categorisation and local gist vision using low-level spatial features.

Object categorisation is a research area with significant challenges, especially in conditions with bad lighting, occlusions, different poses and similar objects. This makes systems that rely on precise information unable to perform efficiently, like a robotic arm that needs to know which objects it can reach. We propose a biologically inspired object detection and categorisation framework that relies on robust low-level object shape. Using only edge conspicuity and disparity features for scene figure-ground segregation and object categorisation, a trained neural network classifier can quickly categorise broad object families and consequently bootstrap a low-level scene gist system. We argue that similar processing is possibly located in the parietal pathway leading to the LIP cortex and, via areas V5/MT and MST, providing useful information to the superior colliculus for eye and head control.

An Energy-Aware Resource Design Model for Constrained Networks

The Internet of Things is expected to incorporate objects and sensor networks of all kinds, and in particular, constrained sensor networks where energy consumption is a critical issue. In order to increase the lifetime of such networks, intelligent and standard-based solutions should be used. Here, we address this challenge through the use of CoRE interfaces for the resource design. These interfaces allow the server side to compose/organize resources and the client side to discover and determine how to consume such resources, besides allowing decisions to be easily integrated into the operation of the network. An energy-aware resource design model is proposed, based on CoRE interfaces, for the design of resources matching client needs. Based on this model, we develop an algorithm that proved to be energy efficient.

Expression-invariant face recognition using a biological disparity energy model

Abstract Biologically–compatible methods are not commonly used for face recognition. Complex computational approaches are preferred and dominate the state of the art. However, we know that the human brain is very efficient at processing faces, without explicitly depending on advanced mathematics. In this paper we focus on evaluating the performance of an expression-invariant face recognition system, which is based on the most widely-accepted biological model of stereo vision: the Disparity Energy Model (DEM), which has been shown to deliver precise but inaccurate results. We show that the DEM can provide 3D disparity maps which are suitable for both identity recognition and verification, even coping with a wide range of facial expressions. We test disparity information, both alone and in combination with luminance data, achieving state-of-the-art results. We also compare DEM results with those obtained by precise and accurate laser range maps, concluding that the differences in performance are very small.

Interoperability in IoT Through the Semantic Profiling of Objects

The emergence of smarter and broader people-oriented IoT applications and services requires interoperability at both data and knowledge levels. However, although some semantic IoT architectures have been proposed, achieving a high degree of interoperability requires dealing with a sea of non-integrated data, scattered across vertical silos. Also, these architectures do not fit into the machine-to-machine requirements, as data annotation has no knowledge on object interactions behind arriving data. This paper presents a vision of how to overcome these issues. More specifically, the semantic profiling of objects, through CoRE related standards, is envisaged as the key for data integration, allowing more powerful data annotation, validation, and reasoning. These are the key blocks for the development of intelligent applications.