Invariant properties of a locally salient dither pattern with a spatial-chromatic histogram

Abstract

Compacted Dither Pattern Code (CDPC) is a recently found feature which is successful in irregular shapes based visual depiction. Locally salient dither pattern feature is an attempt to expand the capability of CDPC for both regular and irregular shape based visual depiction. This paper presents an analysis of rotational and scale invariance property of locally salient dither pattern feature with a two dimensional spatialchromatic histogram, which expands the applicability of the visual feature. Experiments were conducted to exhibit rotational and scale invariance of the feature. These experiments were conducted by combining linear Support Vector Machine (SVM) classifier to the new feature. The experimental results revealed that the locally salient dither pattern feature with the spatialchromatic histogram is rotationally and scale invariant.

Nature Inspired Dimensional Reduction Technique for Fast and Invariant Visual Feature Extraction

“…However, the algorithm was not competent enough to recognize an object with reasonable accuracy. Therefore, in this study, we introduce a completely novel technique to extract SDPF without sacrificing the gained benefits which have been described in [23][24][25].…”

“…The optimal amount of angle bins (k a ) were selected empirically, which will be discussed in the next section. The point allocation to angle bins is done as in (23).…”

A feature clustering approach based on Histogram of Oriented Optical Flow and superpixels

“…where h 1 and h 2 are the two histograms. In this study, the above kernel was implemented as mentioned in [10] using RBF kernel on Bhattachcharya distance between histograms, as shown in (6). k RBF h 1 ,h 2 =exp -d h 1 ,h s (6) where d h 1 ,h s is Bhattachcharya distance between h 1 and h 2 histograms.…”

“…Such assumptions prevent the use of the algorithm for the majority of the videos due to uncertain behaviors of the salient objects. A previous study [4] has showed that a simple clustering technique can be used to cluster salient dither pattern (SDPF) features [6] by using temporal information and dither patterns, for semantic segmentation. However, it requires a human interaction for estimating the number of clusters (K) by inspecting a short sample of the video.…”

“…In this new approach, SLIC superpixel is used to estimate low-level segments in a weighted Lab color space. SDPF is used as the point feature due to its invariant properties [6] and the optical flow is calculated using the pyramidal implementation of Lucas-Kanade optical flow [9]. The segment refining is done by using a short term histogram of oriented optical flow (HOOF) with Radial Basis Function (RBF) kernel on Bhattacharya distance [10] and mean colors of superpixels.…”