Comparative Study of Skin Color based Segmentation Techniques

Abstract

Segmentation is the classification of the input colored image into skin and non-skin pixels based on skin color information. A wide range of applications that require the segmentation process as a preprocessing operation such as computer vision, face/ hand detection and recognition, medical image analysis, and pattern recognition. Color information is one of the simple cues used for detecting skin color, and the use of proper color space to represent color information of an image is a crucial decision. In this literature different segmentation techniques are presented, examples and comparison between the main three based segmentation techniques are given as well. Skin color modeling based statistical model is explained in detail, with discussion the combination with different segmentation techniques. The selection of appropriate segmentation method depends on the application and system environments. The performance of any segmentation algorithm is quantified using some benchmarking such as recall and precision coefficients, or by calculating the percentage of correct and false detection rates according to the complexion of the technique used.

Exploiting Spatial Redundancy of Image Sensor for Motion Robust rPPG

“…Inspired by [14] that using the intensity-normalised rgb and YCrCb to discriminate skin and non-skin regions, this method represents each vector x i with four components: r − g, r − b, Y − Cr and Y − Cb. The OC-SVM is only trained with the first few frames to adapt to the subject skin-tone; then it is used to predict the skin pixels in the subsequent frames, i.e., the pixels with the positive and negative response for f (x) are classified as skin and non-skin pixels respectively.…”

Hand Gesture Recognition: A Literature Review

“…It is the process of dividing the input image (in this case hand gesture image) into regions separated by boundaries [12]. The segmentation process depends on the type of gesture, if it is dynamic gesture then the hand gesture need to be located and tracked [12], if it is static gesture (posture) the input image have to be segmented only.…”

“…The segmentation process depends on the type of gesture, if it is dynamic gesture then the hand gesture need to be located and tracked [12], if it is static gesture (posture) the input image have to be segmented only. The hand should be located firstly, generally a bounding box is used to specify the depending on the skin color [13] and secondly, the hand have to be tracked, for tracking the hand there are two main approaches; either the video is divided into frames and each frame have to be processed alone, in this case the hand frame is treated as a posture and segmented [12], or using some tracking information such as shape, skin color using some tools such as Kalman filter [12].…”

Unsupervised Subject Detection via Remote PPG

“…The key contributions of our work are two-fold: (1) we propose a similarity-based method that exploits the hierarchical voxel-based segmentation and intrinsic properties of human pulse for unsupervised alive subject detection; (2) we develop a spectral analysis algorithm to robustly decompose and update the similarity matrix in the temporal domain, which enables automatic subject number definition. The Voxel-Pulse-Spectral (VPS) method proposed in this work is the first complete solution that uses the pulse for unsupervised alive subject (1) it takes an input video and constructs the hierarchical voxels across the video frames; (2) each voxel simulates an independent pulse-sensor in a parallel pulse extraction process; and (3) all voxels in the hierarchy are pairwisely connected in a similarity matrix based on the measured pulse, while the sparse similar entries denoting the voxel connections are incrementally factorized and fused into a human objectness map.…”

“…However, a common problem faced by these methods is that their trained features are not unique to human beings; any feature that is similar to human skin can be misclassified. Moreover, supervised methods are usually restricted to prior-known samples and tend to fail when unpredictable samples occur, i.e., the Viola-Jones face detector trained with frontal faces cannot locate faces viewed from the side [1], while a skin classifier trained with bright skin fails with dark skin [2]. Fig.…”