A Real-Time Foveated Sensor with Overlapping Receptive Fields

Bolduc, Mélanie; Levine, Martin D.

doi:10.1006/rtim.1996.0056

Cited by 27 publications

(19 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Levine's [16,17] log-polar model of space-variant vision. Logpolar representations have been used to model both cortical magnification [18] and the retina [17].…”

Section: A Space-variant Representation We Use Bolduc Andmentioning

confidence: 99%

Active Object Recognition with a Space-Variant Retina

Kanan

2013

ISRN Machine Vision

View full text Add to dashboard Cite

When independent component analysis (ICA) is applied to color natural images, the representation it learns has spatiochromatic properties similar to the responses of neurons in primary visual cortex. Existing models of ICA have only been applied to pixel patches. This does not take into account the space-variant nature of human vision. To address this, we use the space-variant logpolar transformation to acquire samples from color natural images, and then we apply ICA to the acquired samples. We analyze the spatiochromatic properties of the learned ICA filters. Qualitatively, the model matches the receptive field properties of neurons in primary visual cortex, including exhibiting the same opponent-color structure and a higher density of receptive fields in the foveal region compared to the periphery. We also adopt the "self-taught learning" paradigm from machine learning to assess the model's efficacy at active object and face classification, and the model is competitive with the best approaches in computer vision.

show abstract

“…Levine's [16,17] log-polar model of space-variant vision. Logpolar representations have been used to model both cortical magnification [18] and the retina [17].…”

Section: A Space-variant Representation We Use Bolduc Andmentioning

confidence: 99%

Active Object Recognition with a Space-Variant Retina

Kanan

2013

ISRN Machine Vision

View full text Add to dashboard Cite

show abstract

“…A more flexible solution, to which we subscribe, is the software implementation of the log-polar mapping [26]. The performance through the use of additional hardware is discussed in [27].…”

Section: Introductionmentioning

confidence: 99%

Adaptive Resolution System for Distributed Surveillance

2002

View full text Add to dashboard Cite

“…It turns out that tight frames of compactly supported hybrid splines exist for L 2 ((0, ∞)) and their construction is more natural and more elegant and the analysis is much simpler than those of the Riesz bases. They will be more versatile in applications, which may have wider ramifications, because the processes of foveation, smoothing and compression are not restricted to images but are applicable to a wider class of objects, including those in vision research, 3-D visualization and virtual reality ( [1,2,3,15]), and in the multiresolution processing of geometry information ( [14], [16]). …”

Section: Introductionmentioning

confidence: 99%

Hybrid spline frames

2009

View full text Add to dashboard Cite

Abstract. Using their unitary extension principle, Ron and Shen have constructed a normalized tight frame for L 2 (R) consisting of spline functions with uniform knots. This paper constructs a normalized tight frame for L 2 ((0, ∞)) comprising spline functions with knots on a hybrid of uniform and geometric mesh. The construction is motivated by applications in adaptive approximation using spline functions on a hybrid mesh that admits a natural dyadic multiresolution approximation of L 2 ((0, ∞)) based on dilation and translation.

show abstract

A Real-Time Foveated Sensor with Overlapping Receptive Fields

Cited by 27 publications

References 27 publications

Active Object Recognition with a Space-Variant Retina

Active Object Recognition with a Space-Variant Retina

Adaptive Resolution System for Distributed Surveillance

Hybrid spline frames

Contact Info

Product

Resources

About