摘要
提出一种基于多特征值的服装检测与识别算法。通过人脸检测辅以Canny算子边缘检测实现对服装的精确定位,采用颜色直方图加以最小欧氏距离的组合方式提取特征值,设计贝叶斯分类器,将特征值进行正、负样本分类,最终实现对服装的识别。实验表明,该算法能够有效检测与识别服装,达到预期效果。
Clothing recognition is a research focus in the field of image recognition. This paper studies the clothing detection and recognition technology on the basis of previous studies,and proposes a multi-feature detection and recognition algorithm. Through the face detection combined with Canny edge detection,it achieves precise positioning of clothing,then the characteristic value is extracted by color histogram and minimum Euclidean distance. At the end,design the Bayesian classifier is used to classify the characteristic values into positive and negative samples to identify the clothing. Experiments show that the algorithm can effectively identify the clothing and achieve the desired results.
出处
《实验室研究与探索》
CAS
北大核心
2016年第5期118-122,共5页
Research and Exploration In Laboratory
基金
国家自然科学基金项目(61203246)
