Limits...
Extraction of surface-related features in a recurrent model of V1-V2 interactions.

Weidenbacher U, Neumann H - PLoS ONE (2009)

Bottom Line: The approach is based on feedforward and feedback mechanisms found in visual cortical areas V1 and V2.Unlike previous proposals which treat localized junction configurations as 2D image features, we link them to mechanisms of apparent surface segregation.As a consequence, we demonstrate how junctions can change their perceptual representation depending on the scene context and the spatial configuration of boundary fragments.

View Article: PubMed Central - PubMed

Affiliation: Institute of Neural Information Processing, University of Ulm, Ulm, Germany. ulrich.weidenbacher@uni-ulm.de

ABSTRACT

Background: Humans can effortlessly segment surfaces and objects from two-dimensional (2D) images that are projections of the 3D world. The projection from 3D to 2D leads partially to occlusions of surfaces depending on their position in depth and on viewpoint. One way for the human visual system to infer monocular depth cues could be to extract and interpret occlusions. It has been suggested that the perception of contour junctions, in particular T-junctions, may be used as cue for occlusion of opaque surfaces. Furthermore, X-junctions could be used to signal occlusion of transparent surfaces.

Methodology/principal findings: In this contribution, we propose a neural model that suggests how surface-related cues for occlusion can be extracted from a 2D luminance image. The approach is based on feedforward and feedback mechanisms found in visual cortical areas V1 and V2. In a first step, contours are completed over time by generating groupings of like-oriented contrasts. Few iterations of feedforward and feedback processing lead to a stable representation of completed contours and at the same time to a suppression of image noise. In a second step, contour junctions are localized and read out from the distributed representation of boundary groupings. Moreover, surface-related junctions are made explicit such that they are evaluated to interact as to generate surface-segmentations in static images. In addition, we compare our extracted junction signals with a standard computer vision approach for junction detection to demonstrate that our approach outperforms simple feedforward computation-based approaches.

Conclusions/significance: A model is proposed that uses feedforward and feedback mechanisms to combine contextually relevant features in order to generate consistent boundary groupings of surfaces. Perceptually important junction configurations are robustly extracted from neural representations to signal cues for occlusion and transparency. Unlike previous proposals which treat localized junction configurations as 2D image features, we link them to mechanisms of apparent surface segregation. As a consequence, we demonstrate how junctions can change their perceptual representation depending on the scene context and the spatial configuration of boundary fragments.

Show MeSH

Related in: MedlinePlus

Overview of the core model architecture.The model consists of several stages that were designed to resemble properties of cells found in the early primate visual cortex. Visual input is processed by the hierarchy of different stages from visual area V1 to V2 and vice versa, that is feedforward and feedback. To enhance and complete initial contour signals, recurrent interactions between those two areas are performed iteratively until activities at all stages converge to a stable state. The converged activities can then be read-out from distributed representations to obtain specific maps that signal perceptually important image structures such as completed contours and different types of junction configurations. Such mid-level features provide important cues for occlusion detection or detection of transparencies. In addition, these mid-level features can also play a role in tasks such as border-ownership assignment which perhaps take place in higher visual areas such as V4 or IT.
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC2691604&req=5

pone-0005909-g002: Overview of the core model architecture.The model consists of several stages that were designed to resemble properties of cells found in the early primate visual cortex. Visual input is processed by the hierarchy of different stages from visual area V1 to V2 and vice versa, that is feedforward and feedback. To enhance and complete initial contour signals, recurrent interactions between those two areas are performed iteratively until activities at all stages converge to a stable state. The converged activities can then be read-out from distributed representations to obtain specific maps that signal perceptually important image structures such as completed contours and different types of junction configurations. Such mid-level features provide important cues for occlusion detection or detection of transparencies. In addition, these mid-level features can also play a role in tasks such as border-ownership assignment which perhaps take place in higher visual areas such as V4 or IT.

Mentions: In this section we give a short overview of the proposed model and its components. Our model focuses on the early processing stages of form processing in primate visual cortex, namely cortical areas V1 and V2, and incorporates hierarchical feedforward processing as well as top-down feedback connections to consider the signal flow along the reverse hierarchy processing [6]. An overview of the model architecture is depicted in Figure 2.


Extraction of surface-related features in a recurrent model of V1-V2 interactions.

Weidenbacher U, Neumann H - PLoS ONE (2009)

Overview of the core model architecture.The model consists of several stages that were designed to resemble properties of cells found in the early primate visual cortex. Visual input is processed by the hierarchy of different stages from visual area V1 to V2 and vice versa, that is feedforward and feedback. To enhance and complete initial contour signals, recurrent interactions between those two areas are performed iteratively until activities at all stages converge to a stable state. The converged activities can then be read-out from distributed representations to obtain specific maps that signal perceptually important image structures such as completed contours and different types of junction configurations. Such mid-level features provide important cues for occlusion detection or detection of transparencies. In addition, these mid-level features can also play a role in tasks such as border-ownership assignment which perhaps take place in higher visual areas such as V4 or IT.
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC2691604&req=5

pone-0005909-g002: Overview of the core model architecture.The model consists of several stages that were designed to resemble properties of cells found in the early primate visual cortex. Visual input is processed by the hierarchy of different stages from visual area V1 to V2 and vice versa, that is feedforward and feedback. To enhance and complete initial contour signals, recurrent interactions between those two areas are performed iteratively until activities at all stages converge to a stable state. The converged activities can then be read-out from distributed representations to obtain specific maps that signal perceptually important image structures such as completed contours and different types of junction configurations. Such mid-level features provide important cues for occlusion detection or detection of transparencies. In addition, these mid-level features can also play a role in tasks such as border-ownership assignment which perhaps take place in higher visual areas such as V4 or IT.
Mentions: In this section we give a short overview of the proposed model and its components. Our model focuses on the early processing stages of form processing in primate visual cortex, namely cortical areas V1 and V2, and incorporates hierarchical feedforward processing as well as top-down feedback connections to consider the signal flow along the reverse hierarchy processing [6]. An overview of the model architecture is depicted in Figure 2.

Bottom Line: The approach is based on feedforward and feedback mechanisms found in visual cortical areas V1 and V2.Unlike previous proposals which treat localized junction configurations as 2D image features, we link them to mechanisms of apparent surface segregation.As a consequence, we demonstrate how junctions can change their perceptual representation depending on the scene context and the spatial configuration of boundary fragments.

View Article: PubMed Central - PubMed

Affiliation: Institute of Neural Information Processing, University of Ulm, Ulm, Germany. ulrich.weidenbacher@uni-ulm.de

ABSTRACT

Background: Humans can effortlessly segment surfaces and objects from two-dimensional (2D) images that are projections of the 3D world. The projection from 3D to 2D leads partially to occlusions of surfaces depending on their position in depth and on viewpoint. One way for the human visual system to infer monocular depth cues could be to extract and interpret occlusions. It has been suggested that the perception of contour junctions, in particular T-junctions, may be used as cue for occlusion of opaque surfaces. Furthermore, X-junctions could be used to signal occlusion of transparent surfaces.

Methodology/principal findings: In this contribution, we propose a neural model that suggests how surface-related cues for occlusion can be extracted from a 2D luminance image. The approach is based on feedforward and feedback mechanisms found in visual cortical areas V1 and V2. In a first step, contours are completed over time by generating groupings of like-oriented contrasts. Few iterations of feedforward and feedback processing lead to a stable representation of completed contours and at the same time to a suppression of image noise. In a second step, contour junctions are localized and read out from the distributed representation of boundary groupings. Moreover, surface-related junctions are made explicit such that they are evaluated to interact as to generate surface-segmentations in static images. In addition, we compare our extracted junction signals with a standard computer vision approach for junction detection to demonstrate that our approach outperforms simple feedforward computation-based approaches.

Conclusions/significance: A model is proposed that uses feedforward and feedback mechanisms to combine contextually relevant features in order to generate consistent boundary groupings of surfaces. Perceptually important junction configurations are robustly extracted from neural representations to signal cues for occlusion and transparency. Unlike previous proposals which treat localized junction configurations as 2D image features, we link them to mechanisms of apparent surface segregation. As a consequence, we demonstrate how junctions can change their perceptual representation depending on the scene context and the spatial configuration of boundary fragments.

Show MeSH
Related in: MedlinePlus