Search

Computational models of visual attention have attracted strong interest by accurately predicting how humans deploy attention. However, little research has utilized these models to detect clinical populations whose attention control has been affected by neurological disorders. We designed a framework to decypher disorders from the joint analysis of video and patients' natural eye movement behaviors watch television for 5 minutes. We employ convolutional deep neural networks to extract visual features in real-time at the point of gaze, followed by SVM and Adaboost to classify typically developing children vs. children with fetal alcohol spectrum disorder FASD, who exhibit impaired attentional control. The classifier achieved 74.1% accuracy ROC: 0.82. Our results demonstrate that there is substantial information about attentional control in even very short recordings of natural viewing behavior. Our new method could lead to high-throughput, low-cost screening tools for identifying individuals with deficits in attentional control.

Several salient object detection approaches have been published which have been assessed using different evaluation scores and datasets resulting in discrepancy in model comparison. This calls for a methodological framework to compare existing models and evaluate their pros and cons. We analyze benchmark datasets and scoring techniques and, for the first time, provide a quantitative comparison of 35 state-of-the-art saliency detection models. We find that some models perform consistently better than the others. Saliency models that intend to predict eye fixations perform lower on segmentation datasets compared to salient object detection algorithms. Further, we propose combined models which show that integration of the few best models outperforms all models over other datasets. By analyzing the consistency among the best models and among humans for each scene, we identify the scenes where models or humans fail to detect the most salient object. We highlight the current issues and propose future research directions.

We have created an algorithm to integrate contour elements and find the salience value of them. The algorithm consists of basic long-range orientation specific neural connections as well as a novel group suppression gain control and a fast plasticity term to explain interaction beyond a neurons normal size range. Integration is executed as a series of convolutions on 12 orientation filtered images augmented by the nonlinear fast plasticity and group suppression terms. Testing done on a large number of artificially generated Gabor element contour images shows that the algorithm is effective at finding contour elements within parameters similar to that of human subjects. Testing of real world images yields reasonable results and shows that the algorithm has strong potential for use as an addition to our already existent vision saliency algorithm.

Attentional selection of an object for recognition is often modeled using all-or-nothing switching of neuronal connection pathways from the attended region of the retinal input to the recognition units. However, there is little physiological evidence for such all-or-none modulation in early areas. We present a combined model for spatial attention and object recognition in which the recognition system monitors the entire visual field, but attentional modulation by as little as 20% at a high level is sufficient to recognize multiple objects. To determine the size and shape of the region to be modulated, a rough segmentation is performed, based on pre-attentive features already computed to guide attention. Testing with synthetic and natural stimuli demonstrates that our new approach to attentional selection for recognition yields encouraging results in addition to being biologically plausible.

This paper is a technical description of a new mobile robotics platform specifically designed for the implementation and testing of neuromorphic vision algorithms in unconstrained outdoors environments. The platform is being developed by a team of undergraduate students with graduate supervision and help. Its distinctive features include significant computational power (four 1.1GHz CPUs with gigabit interconnect), high-speed four-wheel-drive chassis, standard Linux operating system, and a comprehensive toolkit of C++ vision classes. The robot is designed with two major goals in mind: real-time operation of sophisticated neuromorphic vision algorithms, and off-the-shelf components to ensure rapid technological evolvability. A preliminary embedded neuromorphic vision architecture that includes attentional, gist/layout, object recognition, and high-level decision subsystems is finally described.

Previous experiments have shown that human attention is influenced by high level task demands. In this paper, we propose an architecture to estimate the task-relevance of attended locations in a scene. We maintain a task graph and compute relevance of fixations using an ontology that contains a description of real world entities and their relationships. Our model guides attention according to a topographic attention guidance map that encodes the bottom-up salience and task-relevance of all locations in the scene. We have demonstrated that our model detects entities that are salient and relevant to the task even on natural cluttered scenes and arbitrary tasks.