摘要
随着智能手机的普及,数码照片的获取变得越来越容易。而社交媒体的盛行,使得大家对分享照片越发热衷。于是更美丽的自拍照就成了大众需求。文中先比较不同人脸检测方法的优缺点,利用肤色模型进行人脸检测,再利用多级中值滤波算法对皮肤进行光滑处理,在不破坏面部细节的情况下能有效减少皱纹和斑点。而利用之前人脸检测对脸部区域的分区,能够很好地对不同区域进行亮度的提升和减少,已达到在视觉上让五官更立体的效果。Open CV对Android平台的支持使得一个好用的图像美化App的实现变得更加容易。文中通过以上人脸美化的方法,编写了一个易用的Android App,可以简单获得一张自动美化过的自拍照。
By the spread of smart phones,digital photos can be gotten more and more easily. People is more and more eager to share photos with the advent of social networking. Then the level-up selfie has become the public＇ s demands. In this paper, the difference between different face detection methods is compared in advantages and disadvantages. The skin color model is used for human face detection, and then the multilevel median filtering algorithm is applied which can effectively reduce wrinkles and spots to smooth the skin in the situation which doesn＇ t destroy the detail of face, use of the partition divided in the former face detection to enhance and reduce the brightness of different regions which want to achieve a goal that makes facial features more three-dimensional in the visual sense. The supporting of OpenCV for the Android platform makes a good face beautification App implementation easier. Through a method above, which beautifies person＇ s face, can develop a convenient Android App which can make users obtain an beautiful selfie simply.
出处
《计算机技术与发展》
2016年第3期9-13,共5页
Computer Technology and Development
基金
国家重大科技专项课题(2009ZX04001-111)
