Omnidirectional Active Vision for Evolutionary Car Driving

Dario Floreano, Mototaka Suzuki, Jacob Van der Blij
2006
Article de conférence

Résumé

We describe a set of simulations to evolve omnidirectional active vision, an artificial retina scanning over images taken via an omnidirectional camera, being applied to a car driving task. While the retina can immediately access features in any direction, it is asked to select behaviorally-relevant features so as to drive the car on the road. Neural controllers which direct both the retinal movement and the system behavior, i.e., the speed and the steering angle of the car, are tested in three different circuits and developed through artificial evolution. We show that the evolved retina moving over the omnidirectional image successfully detects the task-relevant visual features so as to drive the car on the road. Behavioral analysis illustrates its effective strategy in algorithmic, computational, and memory resources.

Official source

https://infoscience.epfl.ch/record/63997?ln=fr

À propos de ce résultat

Cette page est générée automatiquement et peut contenir des informations qui ne sont pas correctes, complètes, à jour ou pertinentes par rapport à votre recherche. Il en va de même pour toutes les autres pages de ce site. Veillez à vérifier les informations auprès des sources officielles de l'EPFL.

Concepts associés

Chargement

Publications associées

Chargement

Parallel and Distributed Neural Models of the Ideomotor Principle: An Investigation of Imitative Cortical Pathways

Aude Billard, Eric Sauser

Humans’ capacity to imitate has been extensively investigated through a wide-range of behavioral and developmental studies. Yet, despite the huge amount of phenomenological evidence gathered, we are still unable to relate this behavioral data to any specific neural substrate. In this paper, we investigate how principles from psychology can be the result of neural computations and therefore attempt to bridge the gap between monkey neurophysiology and human behavioral data, and hence between these two complementary disciplines. Specifically, we address the principle of ideomotor compatibility, by which ‘observing the movements of others influences the quality of one’s own performance’ and develop two neural models which account for a set of related behavioral studies [Brass, M., Bekkering, H., Wohlschla¨ger, A., & Prinz, W. (2000). Compatibility between observed and executed finger movements: comparing symbolic, spatial and imitative cues. Brain and Cognition 44, 124–143]. We show that the ideomotor effect could be the result of two distinct cognitive pathways, which can be modeled by means of biologically plausible neural architectures. Furthermore, we propose a novel behavioral experiment to confirm or refute either of the two model pathways.

2006

Chargement

DeepLabCut: markerless pose estimation of user-defined body parts with deep learning

Pranav Mamidanna, Mackenzie Mathis, Alexander Mathis

Quantifying behavior is crucial for many applications in neuroscience. Videography provides easy methods for the observation and recording of animal behavior in diverse settings, yet extracting particular aspects of a behavior for further analysis can be highly time consuming. In motor control studies, humans or other animals are often marked with reflective markers to assist with computer-based tracking, but markers are intrusive, and the number and location of the markers must be determined a priori. Here we present an efficient method for markerless pose estimation based on transfer learning with deep neural networks that achieves excellent results with minimal training data. We demonstrate the versatility of this framework by tracking various body parts in multiple species across a broad collection of behaviors. Remarkably, even when only a small number of frames are labeled (~200), the algorithm achieves excellent tracking performance on test frames that is comparable to human accuracy.

2018

Chargement

, ,

Quantifying behavior is crucial for many applications in neuroscience. Videography provides easy methods for the observation and recording of animal behavior in diverse settings, yet extracting particular aspects of a behavior for further analysis can be highly time consuming. In motor control studies, humans or other animals are often marked with reflective markers to assist with computer-based tracking, yet markers are intrusive (especially for smaller animals), and the number and location of the markers must be determined a priori. Here, we present a highly efficient method for markerless tracking based on transfer learning with deep neural networks that achieves excellent results with minimal training data. We demonstrate the versatility of this framework by tracking various body parts in a broad collection of experimental settings: mice odor trail-tracking, egg-laying behavior in drosophila, and mouse hand articulation in a skilled forelimb task. For example, during the skilled reaching behavior, individual joints can be automatically tracked (and a confidence score is reported). Remarkably, even when a small number of frames are labeled (

\approx 200

), the algorithm achieves excellent tracking performance on test frames that is comparable to human accuracy.

2018

Chargement

, ,

Quantifying behavior is crucial for many applications in neuroscience. Videography provides easy methods for the observation and recording of animal behavior in diverse settings, yet extracting particular aspects of a behavior for further analysis can be highly time consuming. In motor control studies, humans or other animals are often marked with reflective markers to assist with computer-based tracking, yet markers are intrusive (especially for smaller animals), and the number and location of the markers must be determined a priori. Here, we present a highly efficient method for markerless tracking based on transfer learning with deep neural networks that achieves excellent results with minimal training data. We demonstrate the versatility of this framework by tracking various body parts in a broad collection of experimental settings: mice odor trail-tracking, egg-laying behavior in drosophila, and mouse hand articulation in a skilled forelimb task. For example, during the skilled reaching behavior, individual joints can be automatically tracked (and a confidence score is reported). Remarkably, even when a small number of frames are labeled (

\approx 200

), the algorithm achieves excellent tracking performance on test frames that is comparable to human accuracy.

2018