Independent component analysis of natural image sequences yields spatio-temporal filters similar to simple cells in primary visual cortex

Hateren, van Johannes; Ruderman, Daniel

doi:10.1098/rspb.1998.0577

Cited by 340 publications

(252 citation statements)

References 24 publications

Supporting

Mentioning

240

Contrasting

Unclassified

Order By: Relevance

“…Our main motivation for introducing two types of spatiotemporal RF profiles as shown in Fig. 2 (i.e., stationary and moving envelope) is to explore if there are any significant qualitative differences in the computational properties of one model over the other, to examine the plausibility of suggestions previously made in this context (DeAngelis et al1993a(DeAngelis et al , b, 1995van Hateren and Ruderman 1998) and to check if this is an issue of importance.…”

Section: Spatiotemporal Gabor Filtersmentioning

confidence: 99%

Motion detection, noise reduction, texture suppression, and contour enhancement by spatiotemporal Gabor filters with surround inhibition

Petkov

Subramanian

2007

Biol Cybern

View full text Add to dashboard Cite

We study the orientation and speed tuning properties of spatiotemporal three-dimensional (3D) Gabor and motion energy filters as models of time-dependent receptive fields of simple and complex cells in the primary visual cortex (V1). We augment the motion energy operator with surround suppression to model the inhibitory effect of stimuli outside the classical receptive field. We show that spatiotemporal integration and surround suppression lead to substantial noise reduction. We propose an effective and straightforward motion detection computation that uses the population code of a set of motion energy filters tuned to different velocities. We also show that surround inhibition leads to suppression of texture and thus improves the visibility of object contours and facilitates figure/ground segregation and the detection and recognition of objects.

show abstract

Section: Spatiotemporal Gabor Filtersmentioning

confidence: 99%

Motion detection, noise reduction, texture suppression, and contour enhancement by spatiotemporal Gabor filters with surround inhibition

Petkov

Subramanian

2007

Biol Cybern

View full text Add to dashboard Cite

show abstract

“…The spatiotemporal power spectrum of such passively viewed natural scenes predicts human visual sensitivity (Dong and Atick 1995a;Dong 1997) and response properties of neurons in the lateral geniculate nucleus of the cat (Dong and Atick 1995b). Extending independent component analysis (ICA) from the purely spatial to the spatiotemporal domain on such data predicts the spatio-temporal properties of simple cells (van Hateren and Ruderman 1998). In analogy to sparse coding, different variants of the coding principle ''temporal coherence'', all rooted in Fö ldiak's (1991) trace rule, have been successfully applied to natural stimuli.…”

Section: Introductionmentioning

confidence: 99%

“…Commercial movies are subject to filmmaker bias, and recordings from vehicles (van Hateren and Ruderman 1998) are restricted to a specific type of environment. To overcome these restrictions, Betsch et al (2004) recorded movies using a camera mounted on the head of a freely behaving cat.…”

Section: Introductionmentioning

confidence: 99%

Human eye-head co-ordination in natural exploration

Einhäuser

Schumann

Bardins

et al. 2007

Network: Computation in Neural Systems

View full text Add to dashboard Cite

During natural behavior humans continuously adjust their gaze by moving head and eyes, yielding rich dynamics of the retinal input. Sensory coding models, however, typically assume visual input as smooth or a sequence of static images interleaved by volitional gaze shifts. Are these assumptions valid during free exploration behavior in natural environments? We used an innovative technique to simultaneously record gaze and head movements in humans, who freely explored various environments (forest, train station, apartment). Most movements occur along the cardinal axes, and the predominance of vertical or horizontal movements depends on the environment. Eye and head movements co-occur more frequently than their individual statistics predicts under an independence assumption. The majority of cooccurring movements point in opposite directions, consistent with a gaze-stabilizing role of eye movements. Nevertheless, a substantial fraction of eye movements point in the same direction as co-occurring head movements. Even under the very most conservative assumptions, saccadic eye movements alone cannot account for these synergistic movements. Hence nonsaccadic eye movements that interact synergistically with head movements to adjust gaze cannot be neglected in natural visual input. Natural retinal input is continuously dynamic, and cannot be faithfully modeled as a mere sequence of static frames with interleaved large saccades.

show abstract

“…Recently, there has been a growing interest in learning visual features using biologically-inspired networks, such as, Independent Component Analysis (ICA) [11] and Independent Subspace Analysis [12]. In [13], Le shows that using their learned 3D (Spatiotemporal) filters by ISA, the action recognition performance is comparable to other hand- Fig.…”

Section: Related Workmentioning

confidence: 70%

Complex Events Detection Using Data-Driven Concepts

Yang

Shah

2012

Computer Vision – ECCV 2012

View full text Add to dashboard Cite

Abstract. Automatic event detection in a large collection of unconstrained videos is a challenging and important task. The key issue is to describe long complex video with high level semantic descriptors, which should find the regularity of events in the same category while distinguish those from different categories. This paper proposes a novel unsupervised approach to discover data-driven concepts from multi-modality signals (audio, scene and motion) to describe high level semantics of videos. Our methods consists of three main components: we first learn the low-level features separately from three modalities. Secondly we discover the datadriven concepts based on the statistics of learned features mapped to a low dimensional space using deep belief nets (DBNs). Finally, a compact and robust sparse representation is learned to jointly model the concepts from all three modalities. Extensive experimental results on large in-thewild dataset show that our proposed method significantly outperforms state-of-the-art methods.

show abstract

Independent component analysis of natural image sequences yields spatio-temporal filters similar to simple cells in primary visual cortex

Cited by 340 publications

References 24 publications

Motion detection, noise reduction, texture suppression, and contour enhancement by spatiotemporal Gabor filters with surround inhibition

Motion detection, noise reduction, texture suppression, and contour enhancement by spatiotemporal Gabor filters with surround inhibition

Human eye-head co-ordination in natural exploration

Complex Events Detection Using Data-Driven Concepts

Contact Info

Product

Resources

About