Non-skin Pixels Research Articles

Improving Remote Photoplethysmography Performance through Deep-Learning-Based Real-Time Skin Segmentation Network

In recent years, health-monitoring systems have become increasingly important in the medical and safety fields, including patient and driver monitoring. Remote photoplethysmography is an approach that captures blood flow changes due to cardiac activity by utilizing a camera to measure transmitted or reflected light through the skin, but it has limitations in its sensitivity to changes in illumination and motion. Moreover, remote photoplethysmography signals measured from nonskin regions are unreliable, leading to inaccurate remote photoplethysmography estimation. In this study, we propose Skin-SegNet, a network that minimizes noise factors and improves pulse signal quality through precise skin segmentation. Skin-SegNet separates skin pixels and nonskin pixels, as well as accessories such as glasses and hair, through training on facial structural elements and skin textures. Additionally, Skin-SegNet reduces model parameters using an information blocking decoder and spatial squeeze module, achieving a fast inference time of 15 ms on an Intel i9 CPU. For verification, we evaluated Skin-SegNet using the PURE dataset, which consists of heart rate measurements from various environments. When compared to other skin segmentation methods with similar inference speeds, Skin-SegNet demonstrated a mean absolute percentage error of 1.18%, showing an improvement of approximately 60% compared to the 4.48% error rate of the other methods. The result even exhibits better performance, with only 0.019 million parameters, in comparison to DeepLabV3+, which has 5.22 million model parameters. Consequently, Skin-SegNet is expected to be employed as an effective preprocessing technique for facilitating efficient remote photoplethysmography on low-spec computing devices.

Image Segmentation for Human Skin Detection

Human skin detection is the main task for various human–computer interaction applications. For this, several computer vision-based approaches have been developed in recent years. However, different events and features can interfere in the segmentation process, such as luminosity conditions, skin tones, complex backgrounds, and image capture equipment. In digital imaging, skin segmentation methods can overcome these challenges or at least part of them. However, the images analyzed follow an application-specific pattern. In this paper, we present an approach that uses a set of methods to segment skin and non-skin pixels in images from uncontrolled or unknown environments. Our main result is the ability to segment skin and non-skin pixels in digital images from a non-restrained capture environment. Thus, it overcomes several challenges, such as lighting conditions, compression, and scene complexity. By applying a segmented image examination approach, we determine the proportion of skin pixels present in the image by considering only the objects of interest (i.e., the people). In addition, this segmented analysis can generate independent information regarding each part of the human body. The proposed solution produces a dataset composed of a combination of other datasets present in the literature, which enables the construction of a heterogeneous set of images.

A NOVEL EAR DETECTION TECHNIQUE IN NOISY ENVIRONMENT USING MEAN STRUCTURE SIMILARITY INDEX MEASURE

In this paper, an efficient automated ear detection technique that uses skin colour and templatebased methods is presented. There are two steps involved in the proposed ear detection technique. The first step in this technique is to remove non-skin pixels from the input image, which is called skin segmentation, and the second step is to detect the ear structure using Structure Similarity Index Measure (SSIM). Furthermore, the proposed technique can also be used in noisy environments to detect different shaped ears. This is made possible by creating an ear template using Principal Component Analysis (PCA) and Mean Structure Similarity Index Measure (MSSIM) for template matching and ear localization. An evaluation of the proposed technique is performed using the UND-E Database, which consists of 464 side face images of 114 subjects. It has achieved 97% accuracy in detecting the ear even when the ear image has additive Gaussian, impulsive salt-pepper and multiplicative speckle noise with variance 1/m, with template size of mxn. In particular, the proposed technique is well suited for detecting the ear image of a person moving as it can also detect the ear in motion blurred images.

Human – Ear Recognition Using Scale Invariant Feature Transform

Biometrics techniques are the standard of a wide group of many applications for a human’s identification and verification issues. Because of this reason, a high scale of security needs to search for a new way to identify the person arises. In this paper, establish a human ear recognition system is proposed. This system combines four main phases: ear detection, ear feature extraction, ear recognition, and confirmation. The essential of the proposed system is to divide the ear image into the skin and non-skin pixels using a likelihood skin detector. The likelihood image processes by morphological operations to complete ear regions. Scale-invariant feature transform uses for extracting the fixed features of the ear. Ear recognition includes two modes identification mode and verification mode. Euclidean Distance Measure (EDM) uses for similarity measure between the first image in the database and a new image. According to the three experiments conducted in this paper, the results of the different datasets, the accuracy ratio are 100%, 92%.and 92% respectively.

Image Segmentation for Skin Detection

Human skin detection, which usually performed before image processing, is the method of discovering skin-colored pixels and regions that may be of human faces or limbs in videos or photos. Many computer vision approaches have been developed for skin detection. A skin detector usually transforms a given pixel into a suitable color space and then uses a skin classifier to mark the pixel as a skin or a non-skin pixel. A skin classifier explains the decision boundary of the class of a skin color in the color space based on skin-colored pixels. The purpose of this research is to build a skin detection system that will distinguish between skin and non-skin pixels in colored still pictures. This performed by introducing a metric that measures the distances of pixel colors to skin tones. Results showed that the YCbCr color space performed better skin pixel detection than regular Red Green Blue images due to its isolation of the overall energy of an image in the luminance band. The RGB color space poorly classified images with wooden backgrounds or objects. Then, a histogram-based image segmentation scheme utilized to distinguish between the skin and non-skin pixels. The need for a compact skin model representation should stimulate the development of parametric models of skin detection, which is a future research direction.

Illumination Robust Heart-rate Extraction from Single-wavelength Infrared Camera Using Spatial-channel Expansion.

Heart rate (HR) is one of the most important vital signs for indicating the health condition of a person. Contactless camera-based HR measurement is particularly beneficial for sleep monitoring, as it is comfortable and convenient. However, compared with ambient light, the skin pulsation in near infrared range is much weaker and more susceptible to distortions (e.g. body motion, light changes). In this paper, we propose a method to expand the single-wavelength channel of a near infrared camera to multiple channels for illumination noise reduction, where the channel expansion is performed in the spatial domain using skin and non-skin pixels. The essence is using illumination changes of non-skin pixels to eliminate such a distortion on skin pixels and thus improve pulse extraction. On average, measurement coverage increased from 50% to 83% for the methods of subtraction and Segment Principal Component Analysis (Seg-PCA), and Signal-to-Noise Ratio (SNR) is increased from -8.40 dB to -4.62 dB for the method of Segment Independent Component Analysis (Seg-ICA).

Optimizing Remote Photoplethysmography Using Adaptive Skin Segmentation for Real-Time Heart Rate Monitoring

Choosing a proper Region of Interest (ROI) for Remote Photoplethysmography (rPPG) is essential and a challenging first step, and it has a direct effect on the accuracy and reliability of the overall heart rate (HR) algorithm. Non-skin areas have no contribution to the HR information; however, few works have tackled the issue of non-skin pixels included in the ROI. First, this paper considers improving the quality of the rPPG signal by filtering out non-skin pixels included within the ROI. The feasibility of employing skin segmentation for ROI definition is demonstrated. Then, this technique is compared with our previous real-time rPPG-based method. Moreover, we explore the effect of extracting the HR from three ROIs using signal fusion. Second, we give a comprehensive account of the examined methods in our algorithm for face detection, face tracking, skin detection, and blind signal separation. Finally, we compare our rPPG measurements with ground truth values obtained from a commercial pulse oximeter. Based on the simulation results, the proposed algorrithm significantly improves the quality of the rPPG technique.

Robust cascaded skin detector based on AdaBoost

Skin detection is one of the most important issues of image processing and computer vision. A big concern about skin detection algorithms is being simple while keeping a good accuracy in discriminating skin and non-skin pixels to make the skin detector robust and practicable. This paper proposes a novel and robust skin detector. In this work, statistical information of each pixel and its neighbors is taken into account in order to deal with this concern. A simple algorithm is also presented to reduce the computational complexity of computing the statistical information. In the proposed method, a cascaded classifier by using the AdaBoost algorithm is trained. Finally, two edge detectors are used to make the algorithm more accurate. Moreover, some simple algorithms are used to make the process faster (e.g. algorithms of calculating mean and variance). The performance of the proposed skin detector is evaluated using SFA skin database. In order to illustrate the robustness of the proposed method, the comparisons are made with some popular and newly published skin detectors. The experimental results show that the proposed scheme outperforms other skin detection methods due to high precision and good recall. This method uses a specific way of training AdaBoost in skin detection while having a good accuracy and simplicity compared to other methods.

Robust skin segmentation using color space switching

Skin detection is very popular and has vast applications among researchers in computer vision and human computer interaction. The skin-color changes beyond comparable limits with considerable change in the nature of the light source. Different properties are taken into account when the colors are represented in different color spaces. However, a unique color space has not been found yet to adjust the needs of all illumination changes that can occur to practically similar objects. Therefore a dynamic skin color model must be constructed for robust skin pixel detection, which can cope with natural changes in illumination. This paper purposes that skin detection in a digital color image can be significantly improved by employing automated color space switching. A system with three robust algorithms has been built based on different color spaces towards automatic skin classification in a 2D image. These algorithms are based on the statistical mean of value of the skin pixels in the image. We also take Bayesian approaches to discriminate between skin-alike and non-skin pixels to avoid noise. This work is tested on a set of images which was captured in varying light conditions from highly illuminated to almost dark.

Hybrid Human Skin Detection Using Neural Network and K-Means Clustering Technique

Human skin detection is an essential step in most human detection applications, such as face detection. The performance of any skin detection system depends on assessment of two components: feature extraction and detection method. Skin color is a robust cue used for human skin detection. However, the performance of color-based detection methods is constrained by the overlapping color spaces of skin and non-skin pixels. To increase the accuracy of skin detection, texture features can be exploited as additional cues. In this paper, we propose a hybrid skin detection method based on YIQ color space and the statistical features of skin. A Multilayer Perceptron artificial neural network, which is a universal classifier, is combined with the k-means clustering method to accurately detect skin. The experimental results show that the proposed method can achieve high accuracy with an F1-measure of 87.82% based on images from the ECU database.

Comparative Study of Skin Color Detection and Segmentation in HSV and YCbCr Color Space

Abstract This paper presented a comparative study of human skin color detection HSV and YCbCr color space. Skin color detection is the process of separation between skin and non-skin pixels. It is difficult to develop uniform method for the segmentation or detection of human skin detection because the color tone of human skin is drastically varied for people from one region to another. Literature survey shows that there is a variety of color space is applied for the skin color detection. RGB color space is not preferred for color based detection and color analysis because of mixing of color (chrominance) and intensity (luminance) information and its non uniform characteristics. Luminance and Hue based approaches discriminate color and intensity information even under uneven illumination conditions. Experimental result shows the efficiency of YCbCr color space for the segmentation and detection of skin color in color images.

Research and Improvement on the Algorithm of Face Region Detection Based on Skin Color Model

In face image with complex background, the CbCr skin color region will have offset when considering the illumination change. Therefore, the non-skin color pixels which luminance is less than 80 will be mistaken as skin color pixels and the skin color pixels which luminance is greater than 230 will be mistaken as non-skin color pixels. In order to reduce the misjudgments, an improved skin color model of nonlinear piecewise is put forward in this paper. Firstly, the skin color model of non-piecewise is analyzed and the experimental results show that by this model there is an obvious misjudgment in face detection. Then the skin color model of nonlinear piecewise is mainly analyzed and is demonstrated by mathematics method. A large number of training results show that the skin color model of nonlinear piecewise has better clustering distribution than the skin color model of non-piecewise. At lastly, the face detection algorithm adopting skin color model of nonlinear piecewise is analyzed. The results show that this algorithm is better than the algorithm adopting skin color model of non-piecewise and it makes a good foundation for the application of face image.

Spatial-based skin detection using discriminative skin-presence features

In this paper we propose a new method for skin detection in color images which consists in spatial analysis using the introduced texture-based discriminative skin-presence features. Color-based skin detection has been widely explored and many skin color modeling techniques were developed so far. However, efficacy of the pixel-wise classification is limited due to an overlap between the skin and non-skin pixels reported in many color spaces. To increase the discriminating power of the skin classification schemes, textural and spatial features are often exploited for skin modeling. Our contribution lies in using the proposed discriminative feature space as a domain for spatial analysis of skin pixels. Contrary to existing approaches, we extract the textural features from the skin probability maps rather than from the luminance channel. Presented experimental study confirms that the proposed method outperforms alternative skin detection techniques, which also involve analysis of textural and spatial features.

Development of Multispectral Algorithm for Detection of Cracked Tomatoes

<abstract><title><italic>Abstract.</italic></title> <bold> </bold>Tomato cracks present a potential food safety risk by allowing internalization of pathogens. This article reports a first-step study for the development and evaluation of a simple multispectral algorithm for the detection of defect cracked tomatoes. Analysis of hyperspectral visible/near-infrared tomato images identified two wavebands, 749 nm and 887 nm, for computation of two decision functions to detect defect cracked tomatoes and to differentiate them from normal tomatoes. Detection of defect cracked tomatoes was based on spectral identification of non-skin pixels coupled with quantitative evaluation of the number of non-skin pixels compared to the total image pixels for each tomato. The algorithm successfully classified 94% of the normal tomatoes and 99% of the defect tomatoes used in this study. The results suggest that this multispectral spectral imaging-based method has potential for implementation in line-scan imaging systems as a tool for rapid food inspection on high-speed tomato processing lines.

Comparative Study of Skin Color based Segmentation Techniques

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.

Big skin exposure detection for adult image filtering

With regard to the features of the big skin exposure in the adult images,we propose a hybrid detection approach which is composed of such three parts as color filtering,texture filtering and geometry filtering.A subsection processing skin color model was used to filter the non-skin color pixels and get the candidate skin regions.Then a coarse-degree-based texture filtering was used to filter the pixels with rough texture and skin color tone.At last we filtered the skin-like region(such as deserts or beaches) by fractal dimensions.The experimental results indicate this approach has high detection ratio and high precision.

A dynamic threshold approach for skin tone detection in colour images

This paper presents a novel dynamic threshold approach to discriminate skin pixels and non-skin pixels in colour images. Fixed decision boundaries (or fixed threshold) classification approaches are successfully applied to detect human skin tone in colour images. These fixed thresholds mostly failed in two situations as they only search for a certain skin colour range: any non-skin object may be classified as skin if non-skin objects|s colour values belong to fixed threshold range; any true skin may be mistakenly classified as non-skin if the skin colour values do not belong to fixed threshold range. Therefore in this paper, instead of predefined fixed thresholds, novel online learned dynamic thresholds are used to overcome the above drawbacks. The experimental results show that our method is robust in overcoming these drawbacks.

다층 신경망과 피부색 모델을 이용한 피부 영역 검출

피부색은 자동화된 얼굴 인식을 위한 매우 중요한 정보 중의 하나이다. 본 논문에서는 다층 신경망(Multi-Layer Perceptron)을 이용한 피부 영역 검출 기법을 제안하였다. 제안된 방법은 적응적 조명 보정 기법을 통해 피부색 영역의 검출 성능을 개선하였고, 전처리 필터를 적용하여 피부색이 아닌 영역을 먼저 제거시킴으로써 처리 속도를 향상시켰다. 제안된 방법의 실험 결과 기존의 방법과 비교하여 보다 우수한 검출 결과를 나타냈으며, 처리 속도 또한 약 31~49% 향상시킬 수 있었다. Skin color is a very important information for an automatic face recognition. In this paper, we proposed a skin region extraction method using the MLP(Multi-Layer Perceptron) and skin color model. We use the adaptive lighting compensation technique for improved performance of skin region extraction. Also, using an preprocessing filter, normally large areas of easily distinct non-skin pixels, are eliminated from further processing. Experimental results show that the proposed method has better performance than the conventional methods, and reduces processing time by 31~49% on average.

A new hybrid module for skin detector using fuzzy inference system structure and explicit rules

Skin detection is a popular image processing technique that has been applied in many areas such as video-surveillance, cyber-crime prosecution and face detection. It is also considered as one of the challenging problems in image processing. Despite being a well known technique to detect human appearance within image, it faces a fundamental problem when using colour as cue to detect skin. It is difficult to detect skin when the colour between the skin and the non skin within an image is similar. Therefore in this paper, a new hybrid module between explicit rules and fuzzy inference system structure, based on RGB colour space, is proposed to improve skin detection performance. Using the new hybrid module, we managed to increase the classification reliability when discriminating human skin. The new proposed skin detector depends on subtractive clustering technique, created and trained with training set of skin and non-skin pixels. The proposed system is tested on human images having upright frontal skin with any background. Our proposed system has achieved high detection rates of 87% classification and low false positives when compared with the existing methods. Key words:

Increase reliability for skin detector using backprobgation neural network and heuristic rules based on YCbCr

Skin detection is a popular image processing technique that has been applied in many areas such as video-surveillance, cyber-crime prosecution and unit-spam system. It is also considered as one of the most challenging problems in image processing. Despite being a well known technique in detecting human appearance within image, it faces several drawbacks when using colour as cue to detect skin. First, the difficulty is when the colour between skin and non skin within an image is similar. Second, the skin appearance of humans when exposed under different lighting condition adds the complexity to skin detection. Therefore, this paper proposed a new hybrid module for skin detector using backprobgation neural network and heuristic rules based on YCbCr with the purpose to improve the skin detection performance. Using the new hybrid module, the researcher managed to increase the classification reliability when discriminating human skin colour and regularize the problem when exposed to different lighting conditions. The new proposed skin detector depends on normalization technique to normalize the inputs and targets so that they fall in the interval [-1, 1] and trained with training set of skin and non-skin pixels. Using the techniques discriminates human for images having upright frontal skin with any background achieved high detection rates, scored an 88.5% classification and low false positives when comparing with the previous methods. Key words: Skin detector, backprobgation neural network, explicit rules.