Papers by Claudio CASTELLANOS SÁNCHEZ
On-chip visual perception of motion: A bio-inspired connectionist model on FPGA
Neural Networks, 2005
Visual motion provides useful information to understand the dynamics of a scene to allow intellig... more Visual motion provides useful information to understand the dynamics of a scene to allow intelligent systems interact with their environment. Motion computation is usually restricted by real time requirements that need the design and implementation of specific hardware architectures. In this paper, the design of hardware architecture for a bio-inspired neural model for motion estimation is presented. The motion estimation is based on a strongly localized bio-inspired connectionist model with a particular adaptation of spatio-temporal Gabor-like filtering. The architecture is constituted by three main modules that perform spatial, temporal, and excitatory-inhibitory connectionist processing. The biomimetic architecture is modeled, simulated and validated in VHDL. The synthesis results on a Field Programmable Gate Array (FPGA) device show the potential achievement of real-time performance at an affordable silicon area.
La perception du mouvement visuel modélisé par des modèles connexionnistes fournit différents axe... more La perception du mouvement visuel modélisé par des modèles connexionnistes fournit différents axes de recherche pour le développement de modèles de perception-action en temps réel appliqués à la perception visuelle dynamique du mouvement. Nous présentons dans cet article les bases d'une approche connexionniste bio-inspirée avec une adaptation particulière des filtres de Gabor spatio-temporels que nous utilisons en vue d'une approche plus modulaire et fortement localisée pour la perception visuelle du mouvement. Summary : The visual perception of motion modeled by connectionist models offers various areas of research for the development of models of real-time perception-action applied to the dynamic visual perception of motion. We present in this article the bases of a bio-inspired connectionist approach with a particular adaptation of the spatiotemporal Gabor filters that we use within a more modular and strongly localized approach for the visual perception of motion. Mots clés : Perception du mouvement, réseau de neurones bio-inspirés, flot optique, filtres de Gabor.
Dynamic pursuit with a bio-inspired neural model
Abstract. In this paper we present a bio-inspired connectionist model for visual perception of mo... more Abstract. In this paper we present a bio-inspired connectionist model for visual perception of motion and its pursuit. It is organized in three stages: a causal spatio-temporal filtering of Gabor-like type, an antagonist inhibition mechanism and a densely interconnected neural population. These stages are inspired by the treatment of the primary visual cortex, middle temporal area and superior visual areas. This model has been evaluated on natural image sequences. 1
Snchez, “On-chip visual perception of motion: A bio-inspired connectionist model
Abstract—Visual motion provides useful information to understand the dynamics of a scene to allow... more Abstract—Visual motion provides useful information to understand the dynamics of a scene to allow intelligent systems interact with their environment. Motion computation is usually restricted by real time requirements that need the design and implementation of specific hardware architectures. In this paper, the design of a hardware architecture for a bio-inspired neural model for motion estimation is presented. The motion estimation is based on a strongly localized bio-inspired connectionist model with a particular adaptation of spatio-temporal Gabor-like filtering. The architecture is constituted by three main modules that perform spatial, temporal, and excitatory-inhibitory connectionist processing. The biomimetic architecture is modeled, simulated and validated in VHDL. The synthesis results on a Field Programmable Gate Array (FPGA) device show the potential achievement of real-time performance at an affordable silicon area.
Memorias asociativas morfológicas
Neuromimetic motion indicator for visual perception
This paper presents a bio-inspired model for visual perception of motion through its principal in... more This paper presents a bio-inspired model for visual perception of motion through its principal indicator : the neuromimetic motion indicator (NMI). This indicator emerges out of the mechanism of antagonist interactions (MAI) where an architecture of oriented columns, local and distributed interactions of the neurons in the primary visual cortex (V1) is operated. The NMI indicator classifies the motion in : null motion and motion, and it estimates the number of moving objects in the scene.
Towards bio-inspired neural networks for visual perception of motion
ABSTRACT
Towards a Bio-Inspired Approach to Clustering Facial Gesture in Suspect/Non-Suspect Persons
Proceedings of the International Conference on Computer Vision Theory and Applications, 2012
Visual Cortex - Current Status and Perspectives, 2012
Brain, Vision and AI, 2008
Le recherche en connexionnisme est en face des fortes contraintes d'autonomie en robotique no... more Le recherche en connexionnisme est en face des fortes contraintes d'autonomie en robotique notamment issues des tâches perceptives visuelles qui n'ont pas encore ete satisfaites par les modeles neuronaux utilises jusqu'a aujourd'hui, surtout celles liees a la perception du mouvement. Dans ce cadre, nous proposons un modele connexionniste neuromimetique permettant de traiter un ensemble de tâches de perception visuelle dynamique (robot/objets environnants en mouvement). Cela implique le compromis de maintenir la satisfaction des contraintes (bio-inspiration, traitements locaux massivement distribue��s en vue d'une implantation ulterieure temps-reel embarquee sur circuit FPGA, Field Programmable Gate Array) au cœur de la definition du modele propose, malgre la complexite des tâches en jeu. Le modele connexionniste propose pour la perception visuelle du mouvement est constitue de trois modules : le premier opere un filtrage spatio-temporel causal issu des filtres de...
Bio-inspired Architecture for Human Detection
In this paper we propose a bio-inspired architecture to detect, describe and distinguish objects ... more In this paper we propose a bio-inspired architecture to detect, describe and distinguish objects in motion. By using neuronal and physiological mechanisms in primary visual cortex (V1), middle temporal (MT) and inferotemporal (IT) areas we can start isolating the objects from their environment; then, track, label and distinguish the humans from non-human figures in motion and finally, represent the person’s silhouette to get a better understanding of the body structure.
A Bio-inspired Connectionist Architecture for Visual Classification of Moving Objects
In this paper we propose a bio-inspired connectionist architecture for visual classification of m... more In this paper we propose a bio-inspired connectionist architecture for visual classification of moving objects in articulated and non-articulated ones. It is based on a previous model called CONEPVIM, which uses the behaviour of simples cells in the visual primary cortex (V1) to estimate motion. Besides, on general new ideas on the area of neurophysiology for detecting the articulated moving objects and understanding the processing of the visual information from the earlier vision. We tested our architecture with real images without illumination constraints and any previous knowledge.
Digital implementation of a bio-inspired neural model for motion estimation
... Digital Implementation of a Bio-inspired Neural Model for Motion Estimation CesarTorres-Huitz... more ... Digital Implementation of a Bio-inspired Neural Model for Motion Estimation CesarTorres-Huitzil LORIA-INRIA Lorraine Nancy, France E-mail: torres@loria.fr Bernard Girau LORIA-INRIA Lorraine Nancy, France E-mail: giraugloria.fr ...
Neuromimetic motion indicator for visual perception
Dynamic Pursuit with a Bio-inspired Neural Model
In this paper we present a bio-inspired connectionist model for visual perception of motion and i... more In this paper we present a bio-inspired connectionist model for visual perception of motion and its pursuit. It is organized in three stages: a causal spatio-temporal filtering of Gabor-like type, an antagonist inhibition mechanism and a densely interconnected neural population. These stages are inspired by the neural treatment and the interactions of the primary visual cortex, middle temporal area and superior visual areas. This model has been evaluated on natural image sequences.
Neuromimetic Indicators for Visual Perception of Motion
This paper presents three neuromimetic indicators for the visual perception of motion. They estim... more This paper presents three neuromimetic indicators for the visual perception of motion. They estimate the motion, the speed and the direction. All of them emerge from the first two stages in the Castellanos model [1], where a causal spatio-temporal filtering of Gabor-like type captures the oriented contrast and an antagonist inhibition mechanism estimates the motion. These neuromimetic indicators have been evaluated on sequences of natural and synthetic images.
Bio-inspired Architecture for Visual Recognition of Humans Walking
In this paper we propose a bio-inspired architecture for visual recognition of humans at walking ... more In this paper we propose a bio-inspired architecture for visual recognition of humans at walking and objects that can be humans but do not describe a gait like humans at walking, based on the behaviour of simples cells in the human primary visual cortex. This architecture was tested with real sequences of images acquired in natural environments. The results show the flexibility of our propose since it helps to distinguish between these two types of moving objects, even in unknown scene conditions (bright, or background motion).
Bio-inspired Connectionist Architecture for Visual Detection and Refinement of Shapes
In this paper we propose a bio-inspired architecture for the visual reconstruction of silhouettes... more In this paper we propose a bio-inspired architecture for the visual reconstruction of silhouettes of moving objects, based on the behaviour of simple cells, complex cells and the Long-Range interactions of these neurons present in the primary visual cortex of the primates. This architecture was tested with real sequences of images acquired in natural environments. The results combined with our previous results show the flexibility of our propose since it allows not only to reconstruct the silhouettes of objects in general, but also, allows to distinguish between different types of objects in motion. This distinction is necessary since our future objective is the identification of people by their gait.
Bio-inspired Clustering of Moving Objects
Page 1. Bio-inspired clustering of moving objects Ivonne Maricela Avila-Mora Laboratory of Inform... more Page 1. Bio-inspired clustering of moving objects Ivonne Maricela Avila-Mora Laboratory of Information Technologies Cinvestav - Tamaulipas, México Email: iavila@tamps.cinvestav.mx Claudio Castellanos-Sánchez Laboratory ...
Uploads
Papers by Claudio CASTELLANOS SÁNCHEZ