Computer Engineering and Applications ›› 2019, Vol. 55 ›› Issue (22): 187-194.DOI: 10.3778/j.issn.1002-8331.1808-0108

Previous Articles     Next Articles

Face Sketch Synthesis Method Combining pHash and Sparse Coding

ZHANG Hua, CAO Lin   

  1. Department of Telecommunication Engineering, Beijing Information Science and Technology University, Beijing 100101, China
  • Online:2019-11-15 Published:2019-11-13



  1. 北京信息科技大学 通信工程系,北京 100101

Abstract: In order to solve the problems of blurred details and low definition of existing face sketch synthesis methods, this paper presents a face sketch synthesis method based on Perceptual Hash(pHash) algorithm and Sparse Coding(SC). Firstly, according to the information entropy of the image, the photo-sketch pairs are adaptively divided into blocks. Next, the Hash fingerprints of the large image blocks are calculated by pHash algorithm, and the small image blocks are sparsely coded. Then, the [K] initial candidate photo blocks which are most similar to the test blocks are selected and the corresponding sketch blocks are obtained. Finally, the second sparse coding method is introduced to synthesize the final sketch block, and then the whole face sketch image is got. The validity of the algorithm is verified by existing face databases, which can be used for sketch face synthesis after optimization.

Key words: face sketch synthesis, image information entropy, adaptive block, perceptual hash algorithm, sparse coding

摘要: 为解决已有素描人脸合成方法存在的细节模糊和清晰度低的问题,提出一种感知哈希算法(Perceptual Hash,pHash)与稀疏编码(Sparse Coding,SC)相结合的素描人脸合成方法。首先根据图像的信息熵对人脸照片-素描对进行自适应分块处理,利用感知哈希算法计算出大图像块的哈希指纹,并对小图像块进行稀疏编码;然后选取与测试照片块最相似的[K]个初始候选照片块,得到与之对应的素描块;最后引入二次稀疏编码方法,合成最终的素描块,进而合成整幅素描人脸图像。利用现有的人脸数据库验证了算法的有效性,该算法经优化后可用于素描人脸合成。

关键词: 素描人脸合成, 图像信息熵, 自适应分块, 感知哈希算法, 稀疏编码