UD Vision Lab

Face Sketch Recognition

Face sketch recognition aims at recognizing faces from hand drawn sketches. A typical application of face sketch recognition is in criminal investigation to locate criminals involved. There are several challenges associated with face sketch recognition. The sketches are highly dependent on the artist’s rendering style. It is unlikely for the artists to render all the features correctly, as the sketches are drawn purely based on the descriptions of witnesses or victims. Since a sketch and photo are not really identical, conventional face recognition methods cannot be applied directly to the sketch. We have devised a new methodology for face sketch recognition. The results on CUHK (Chinese University of Hong Kong) student dataset show the effectiveness of our approach in face sketch recognition.

Results on the CUHK (Chinese University of Hong Kong) student dataset:
The database has 188 faces [1]. 88 out of this forms the training set and the remaining 100 forms test set. Some sample images from CUHK student dataset are shown in figure 1. Our method is compared with some conventional methods [2]. These results are shown in figure 2.

Face Recognition Sample Data

Face Recognition Graph

References:

1. X. Wang and X. Tang, "Face Photo-Sketch Synthesis and Recognition," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 31, pp. 1955-1967, 2009.
2. X. Tang, and X. Wang, "Face Photo Recognition Using Sketch," Proceedings of IEEE International Conference on Image Processing, vol. 1, pp. 257-260, 2002.