The Landscape of Research on Skin Detectors: Coherent Taxonomy, Open Challenges, Motivations, Recommendations and Statistical Analysis, Future Directions

Chyad, M. A.; Alsattar, Hassan A.; Zaidan, B. B.; Zaidan, A. A.; Shafeey, Ghailan A. Al

doi:10.1109/access.2019.2924989

Cited by 14 publications

(6 citation statements)

References 168 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Color-based skin detection methods are still the mainstream in time-sensitive applications [46].Albert, the essential importance of the skin classifiers, a large number of similar studies and reviews keep popping out without pointing to a final perfect solution [47]. Regarding the wide variety of color spaces, skin and camera modeling, and illumination, each method has advantageous usages, and no single method can conquer other methods [48].…”

Section: B Relation To Related Researchmentioning

confidence: 99%

High-Reliability Non-Contact Photoplethysmography Imaging for Newborn Care by a Generative Artificial Intelligence

Chen,

Lin,

Yang

et al. 2023

IEEE Access

View full text Add to dashboard Cite

Long-term wiring on a newborn patient could be a disguise scene for parents. Unobtrusive and reliable monitoring without wiring can be a euphoric alternative for newborns and parents in obstetrics and gynecology (OB/GYN) incubation rooms. However, reliable and continuous non-contact surveillance in an incubation room is challenging. Therefore, a novel photoplethysmography imaging (PPGi) is developed specifically for baby skins through predictive adversarial adaptation and risk-sensitive generative synchronizer. Our artificial intelligence approach does not take blind guesses from input-output pairs. We apply an intelligent step to decouple the influence of fluctuated illumination through a generative algorithm of artificial intelligence. To boost skin detection performance, we capture those pixels with periodic variations and maximize the coherence of the extraction algorithm by the generative synchronizer. The periodic variations are matched by a synthesized pulse from the output PPGi signals through the control of a risksensitive filter to not over-compensate the illuminate variation. Based on the sensed pulsation, we synthesize the corresponding pulsation signals on the flight to identify the living skin in a spatiotemporal image sequence. We find that our skin classifier in risk-sensitive generative synchronizer effectively improves the quality of the resulting non-contact PPGi signal. Our algorithm produces substantial accuracy in the performance of PPGi reconstruction in the critical environment of newborn care. In the limited illustration of the incubation room, our non-contact PPGi can still achieve an average accuracy of 96.62%.

show abstract

Section: B Relation To Related Researchmentioning

confidence: 99%

High-Reliability Non-Contact Photoplethysmography Imaging for Newborn Care by a Generative Artificial Intelligence

Chen,

Lin,

Yang

et al. 2023

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Although the word vectors processed by One-hot coding can be directly input into the model, there are two problems in the word vectors after One-hot coding: first, the dimension of the word vectors is large and the sparsity is high, which will lead to the excessive dimension of the vocabulary and bring dimensional disaster. Secondly, the word vector encoded by One-HOT cannot represent the similarity between words, ignoring the semantic information of sentences [14].…”

Section: Figure 1 Basic Process Of Web Crawlermentioning

confidence: 99%

Deep Learning in Urban Tourism Route Decision Suort

2021

View full text Add to dashboard Cite

Users' demand for personalized travel services is becoming stronger and stronger. Providing more intelligent travel suggestions to users has become a hot topic in academia and industry. This paper mainly studies the alication of deep learning in urban tourism route decision suort. Due to the advantages of CNN network in text feature extraction, the design uses CNN network to process user comment information and tourism service project comment information respectively. The network is designed with four layers: input layer, convolution layer, pooling layer, and fully connected layer. A set of tourism service recommendation model based on deep learning is constructed and implemented. Through the experimental comparison with the common traditional recommendation methods, it can be found from the experimental results that the proposed model method is significantly better than the traditional recommendation methods, which verifies the superiority of the model.

show abstract

“…These approaches determine the nude body through detecting the human skin [11], [20], and then define appropriate thresholds of human skin regions in the entire image to recognize the pornography [21]. The human skin-based approaches are highly dependent on the detection of human skin color [22]. For instance, many researchers focused on the color space model to improve the accuracy of the detection of human skin [12], [21]- [24].…”

Section: Related Workmentioning

confidence: 99%

A Pornographic Images Recognition Model based on Deep One-Class Classification With Visual Attention Mechanism

Chen

Liang

et al. 2020

IEEE Access

View full text Add to dashboard Cite

The existing approaches usually treat the recognition of pornographic images as a binary or multi-class classification task, which is realized by extracting a variety of features. However, these approaches ignore the problem of infinite kinds in the negative samples and lose sight of the impact of uncertainty in classification tasks, resulting in inadequate negative data sets and incorrect recognition of samples that are not in the training set. In order to address this challenge, this paper proposes a method named Deep One-Class with Attention for Pornography (DOCAPorn) that recognizes the pornographic images through the one-class classification model based on neural networks and introduces the visual attention mechanism to enhance the performance of recognition. Moreover, since the existing approaches based on deep convolutional neural networks (CNNs) require a fixed-size (e.g., 224 × 224) input image, the geometric distortion caused by image scaling is not considerd in the existing approaches, which reduces the pornographic image recognizition accuracy. In order to solve this issue, this paper proposes the Scale Constraint Pooling (SCP) that converts the inputs of different dimensions into outputs of the same dimension. In addition, all the existing approaches ignore the adversarial attacks in the field of pornographic image recognition. In order to deal with this problem, the paper proposes the Preprocessing for Compressing and Reconstructing (PreCR), a pre-processing approach that reduces the subtle perturbation through compressing the images and then reconstructs the purified image for recognition. The proposed approach is verified by conducting comparative experiments using custom datasets. The experimental results showed that we achieved an accuracy of 98.419% on our dataset. In addition, the proposed approach yielded a recognition accuracy of 95.632% on the NPDI dataset. Furthermore, the obtained results demonstrate that the proposed approach not only effectively recognizes pornographic images but also effectively defends the adversarial attacks.

show abstract

The Landscape of Research on Skin Detectors: Coherent Taxonomy, Open Challenges, Motivations, Recommendations and Statistical Analysis, Future Directions

Cited by 14 publications

References 168 publications

High-Reliability Non-Contact Photoplethysmography Imaging for Newborn Care by a Generative Artificial Intelligence

High-Reliability Non-Contact Photoplethysmography Imaging for Newborn Care by a Generative Artificial Intelligence

Deep Learning in Urban Tourism Route Decision Suort

A Pornographic Images Recognition Model based on Deep One-Class Classification With Visual Attention Mechanism

Contact Info

Product

Resources

About