Perceptually Motivated Image Features Using Contours

Abstract

Abstract-Dong et al. examined the ability of 51 computational feature sets to estimate human perceptual texture similarity, however, none performed well for this task. While it is well-known that the human visual system is extremely adept at exploiting longer-range aperiodic (and periodic) "contour" characteristics in images, none of the investigated feature sets exploit higher order statistics (HOS) over larger image regions (>19×19 pixels). We therefore hypothesise that long-range HOS, in the form of contour data, are useful for perceptual texture similarity estimation.We present the results of a psychophysical experiment that shows that contour data are more important, than local image patches, or global 2nd-order data, to human observers for this task.Inspired by this finding, we propose a set of perceptually motivated image features (PMIF) that encode the long-range HOS computed from spatial and angular distributions of contour segments. We use two perceptual texture similarity estimation tasks to compare PMIF against the 51 feature sets referred to above and four commonly used contour representations. This new feature set is also examined in the context of two additional tasks: sketch-based image retrieval and natural scene recognition. The results show that the proposed feature set performs better, or at least comparably to, all the other feature sets. We attribute this promising performance to the fact that the proposed feature set exploits both short-range and long-range HOS.

The Visual Word Booster: A Spatial Layout of Words Descriptor Exploiting Contour Cues

“…However, these elements often contain more than one segment and thereby cannot be used to represent the shape of a single contour. Compared with the PMIF descriptor [18] based on contours, this paper addresses its open issues by representing local contour segments in a more powerful way via conton learning, and modelling the global spatial layout of words in longer ranges.…”

“…Inspired by the importance of contour cues to human visual perception of imagery [15], [18], [23], [51], [68], we introduce a global image descriptor (see Fig. 1 for pipeline) which exploits the spatial layout of words (SLoW) based on contours.…”

“…In this situation, the exploitation of the contour cues in images provides a possible solution to encoding global spatial layouts [26][27], [55]. It has been highlighted in the literature [15], [18], [23], [51], [68] that contour cues are important to human visual perception of imagery. Dong and Chanter [18] examined the importance of contour cues to texture perception.…”

“…It has been highlighted in the literature [15], [18], [23], [51], [68] that contour cues are important to human visual perception of imagery. Dong and Chanter [18] examined the importance of contour cues to texture perception. On the basis of contours, they proposed a descriptor, namely, Perceptually Motivated Image Features (PMIF).…”

A Novel Methodology for Efficient Retrievals of Sketch Based Images using Local Binary Patterns and Generalized Gamma Mixture Models (GGMM)

“…The area of Content Based Image Retrieval has given a note worthy stride in the area of image retrievals and the main advantage with this mechanism is that based on content effective retrievals from massive datasets are made possible. This content can be either the visual features ,textual features, shape features or any other feature that helps in proper identification of image under studying and thereby helping towards efficient retrievals [1] [2][3][4] [5].With this features many applications have been developed and are being utilized in real time situations including the domains of health care, medical, E-business, etc., [6] [7] [8]. Many models are thus subjected to using both parametric model based approaches and non parametric model based approaches.…”