敏感图片自动检测系统的研究

设计并实现了一个敏感图片自动识别系统。该系统首先利用肤色模型从敏感图片中检测出肤色区域,再从图片肤色区域中,提取大量经验特征表示图像内容,然后采用Haar特征算法设计一个检测敏感部位对象的分类器,最后通过敏感部位的几何特征逐级识别正常图片与敏感图片。该检测系统具有良好的系统性能。     

基于GPU的皮肤变形算法

为提高基于实例的皮肤变形算法的效率和速度,对加权姿态空间变形(WPSD)等算法进行了研究.WPSD中需要的逐顶点计算可以在GPU上以单指令流多数据流(SIMD)方式并行执行.并行顶点计算一般在GPU顶点处理器中执行,但是更进一步的并行操作是可以由片段处理器来获取到的.提出了一种基于GPU片段处理器实现的并行变形算法.每个顶点的关节权值可以由样本姿态自动的计算出,因此减少了手动操作,提高了WPSD等皮肤变形算法的质量.     

基于肤色模型和梯度算子的正面人脸轮廓提取

文章提出了一种新的正面人脸封闭轮廓提取方法,首先采用肤色模型对肤色区域进行初定位,然后经过筛选去除非人脸区域,最后提取人脸初步轮廓,并结合梯度算子提取下巴轮廓,进而得到一个连续、封闭的人脸轮廓.该算法基本克服了下巴轮廓难以从颈部提取的问题,满足了人脸轮廓特征提取的要求,同时具有较高的时效性.     

基于肤色检测技术的手势分割

手语是聋哑人使用的语言。在手势输入设备发展二十多年以后,许多人仍然发现在与聋哑人的互动交流中存在着一定的困难。因此应该努力使计算机适应自然的交流方式。文中的目的是利用肤色检测程序从整个图像上获取有意义的手势区域,并处理不同室内背景和照明情况下的手势,为进一步的手语识别做准备。对几个静态手势的检测过程进行了实验评估,并就结果进行了讨论。     

基于肤色和Adaboost算法的人脸检测研究

人脸检测是计算机视觉和人工智能领域中的一项富有挑战性的工作,在虚拟现实、人机交互等很多领域都有广泛的应用。研究了基于Adaboost的人脸检测,并提出了肤色与Adaboost算法相结合的人脸检测方法。对输人的彩色图像进行从RGB空间到YCrCb空间的转换,再结合形态学等方法进行区域肤色分割,排除背景干扰,然后用Adaboost算法对可能区域进行检测,得到人脸位置。实验表明,该方法有较高的准确性和鲁棒性,可以得到满意的检测效果。     

利用物理模型合成皮肤图像

为了获得各种丰富和逼真的皮肤图像,在对皮肤结构和皮肤组成进行分析的基础上,提出了一种基于物理模型的鲁棒算法对输入的单张皮肤图像自动进行色素分离,并在此基础上实现了全局性和局部性的合成。输入的图像首先被分成N个子区域,并依次在每个子区域利用ICA算法提取分离向量,之后对得到的分离向量进行合理性验证。如果分离向量不合理,则丢弃,在下一个子区域重复上述过程。最后,从所有的合理分离向量中得到最终的分离向量,分离得到黑色素和血色素的密度分布图。通过改变色素的密度可以进行全局性合成,也可以利用密度信息选取感兴趣区域进行局部性合成。实验结果表明,该算法非常有效,可以对人体皮肤图像进行"电子化妆",重新生成更多逼真的合成图像。     

基于黑曲霉固态发酵的绿豆皮降解及理化特性提升机理

本试验研究了黑曲霉固态发酵对绿豆皮特性的影响,并对降解机理进行探究,结果表明黑曲霉固态发酵能有效提升绿豆皮制粉特性和功能特性,所得绿豆皮粉体积平均径减小8.53倍,比表面积增大32.46倍。阳离子交换能力、胆酸钠吸附能力及葡萄糖吸附能力均有明显提升,依次提升1.11倍、1.18倍和1.17倍。黑曲霉对绿豆皮的降解,先降解木质素与半纤维素,再降解纤维素。主要通过破坏纤维素、半纤维素及木质素的化学键,降解长链结构为短链结构,破坏结晶区结构,降低相对结晶度,进而破坏绿豆皮结构,使其变得疏松多孔,让绿豆皮易被粉碎为粒径更小、比表面积更大的粉体,具有更好的物质交换效率,提升绿豆皮的理化特性。     

Latch Applicator for Efficient Delivery of Dissolving Microneedles Based on Rapid Release of Elastic Strain Energy by Thumb Force (Adv. Funct. Mater. 11/2023)

Dissolving microneedle (DMN) is an attractive alternative to parenteral and enteral drug administration owing to its painless self-administration and safety due to non-generation of medical waste. For reproducible and efficient DMN administration, various DMN application methods, such as weights, springs, and electromagnetic devices, have been studied. However, these applicators have complex structures that are complicated to use and high production costs. In this study, a latch applicator that consists of only simple plastic parts and operates via thumb force without any external complex device is developed. Protrusion-shaped latches and impact distances are designed to accumulate thumb force energy through elastic deformation and to control impact velocity. The optimized latch applicator with a pressing force of 25 N and an impact velocity of 5.9 m s⁻¹ fully inserts the drug-loaded tip of the two-layered DMN into the skin. In an ovalbumin immunization test, DMN with the latch applicator shows a significantly higher IgG antibody production rate than that of intramuscular injection. The latch applicator, which provides effective DMN insertion and a competitive price compared with conventional syringes, has great potential to improve delivery of drugs, including vaccines.     

Optimal Artificial Intelligence Based Automated Skin Lesion Detection and Classification Model

Skin lesions have become a critical illness worldwide, and the earlier identification of skin lesions using dermoscopic images can raise the survival rate. Classification of the skin lesion from those dermoscopic images will be a tedious task. The accuracy of the classification of skin lesions is improved by the use of deep learning models. Recently, convolutional neural networks (CNN) have been established in this domain, and their techniques are extremely established for feature extraction, leading to enhanced classification. With this motivation, this study focuses on the design of artificial intelligence (AI) based solutions, particularly deep learning (DL) algorithms, to distinguish malignant skin lesions from benign lesions in dermoscopic images. This study presents an automated skin lesion detection and classification technique utilizing optimized stacked sparse autoencoder (OSSAE) based feature extractor with backpropagation neural network (BPNN), named the OSSAE-BPNN technique. The proposed technique contains a multi-level thresholding based segmentation technique for detecting the affected lesion region. In addition, the OSSAE based feature extractor and BPNN based classifier are employed for skin lesion diagnosis. Moreover, the parameter tuning of the SSAE model is carried out by the use of sea gull optimization (SGO) algorithm. To showcase the enhanced outcomes of the OSSAE-BPNN model, a comprehensive experimental analysis is performed on the benchmark dataset. The experimental findings demonstrated that the OSSAE-BPNN approach outperformed other current strategies in terms of several assessment metrics.