Noise-Against Skeleton Extraction Framework and Application on Hand Gesture Recognition

Abstract

Extracting stable skeletons from noisy images is a challenging problem since the skeletonization method is prone to be affected by inner and border noise. Although many methods have been proposed in the past for increasing the antinoise ability of skeletonization methods, most of them either only overcome border noise or, at the cost of lost topology, degrade the effects of two noises. In this paper, we propose a skeleton extraction framework to enhance the robustness of the existing skeletonization method against both inner and border noise. In our approach, we first use the different scales of Gaussian filters to smooth the input image and obtain multiple representations. Then, binarization and skeletonization were performed to produce a series of binary images and a series of skeletal images. Next, we use our measure on these binary and skeletal images to find the most suitable skeleton. Since our measure considers both the skeleton image changes and binary image changes caused by using a filter, the selected skeleton is sufficiently robust and has all the necessary skeletal branches. The inner noise experiment and border noise experiment are conducted for comparison. From the perspective of the measure of the rate of variation in the skeleton, the proposed framework can reduce the inner noise by approximately 92% and the border noise by approximately 40%. In addition, the experiment on static hand gesture recognition has demonstrated that the introduction of our framework can increase up to 11% mean recognition accuracy.INDEX TERMS Gaussian filter, noise against, skeleton extraction framework, static hand gesture recognition.

Survey of Noise-Against Techniques for Extracting Stable Skeleton

“…To better understand denoising techniques that mentioned in previous section, we have conducted both the inner noise experiment and border noise experiment on method OPTA [27],MOPCA(skeletonization-based denoising) [7], OPCA+DCEM(pruning-based denoising) [22], and OPCA+ATMF(scale-space-based denoising) [26]. The OPTA method is a standard skeletonization, which is sensitive to both of two noise.…”

“…Scale-space-based denoising approaches [23]- [26] adopt scale space filters before the binarization stage to smooth the potential noise. These kinds of methods can simultaneously suppress both inner noise and border noise.…”

“…Then, they used their proposed sensitive measure to evaluate these skeletons and from which the skeleton with the lowest score is viewed as the stable skeleton. Ma [26] recently proposed an improved version of Chatbri's method and applied their method to static hand gesture recognition. Their modification is major on the sensitive measure, in which more information is included.…”

Skeleton extraction of hard‐pen regular script based on stroke characterization and ambiguous zone detection

“…Consequently, there is a risk of shorter burrs being incorrectly identified as stroke branches and not being effectively removed. Ma et al [17] proposed an automatic pruning method based on Bai's pruning algorithm [14]. The method utilizes DCE to extract salient points from the graphic contour, and pairs and sorts the endpoints and salient points.…”

“…According to the algorithm for deburring in SCAD, the stroke characteristics of hard-pen regular script are analyzed first, and then different heuristic rules are set for different kinds of burrs to detect burrs in a targeted way. The automatic pruning method [17] utilizes DCE to extract salient points from the graphic contour. The endpoints and salient points are paired and sorted, and the endpoints and their skeleton branches are sequentially selected to form the reconstructed image.…”