An Empirical Investigation on the Effect of Shape Exaggeration in Face Sketch to Photo Matching
Samsul Setumin, Abduljalil Radman, Shahrel Azmin Suandi · 2019
Facial sketch to photo retrieval system has useful application in forensic investigation. The main problem identified in the system is that the sketch and photo are from different modality. In order to address this problem, intra-modality and inter-modality matching approaches have been introduced by many researchers. However, the proposed methods mostly ignore the effects of shape exaggeration of the facial sketch. In this paper, we attempt to investigate empirically and highlight the importance of considering the shape exaggeration effects in any of the two aforementioned matching approaches. This investigation is limited to the Chinese University of Hong Kong (CUHK) Face Sketch Database (CUFS) and CUHK Face Sketch FERET Database (CUFSF) datasets. The results demonstrate that the shape exaggeration effects may reduce retrieval rate accuracy.