Open Access Open Access  Restricted Access Subscription Access

Exploring the Challenges and Opportunities of Deep Learning Models in Criminal Face Sketch Analysis: From Summary to Identification

Kannan N, Jeeva S, Rosenpranav K S, Mohan Raja V, Dr. V. Kamalaveni

Abstract


Generating accurate criminal face sketches from witness descriptions is essential in modern law enforcement. Deep learning (DL) has advanced this process but faces limitations, including large dataset requirements, poor generalization to photos, and high computational costs. This paper reviews DL-based face sketch synthesis, highlighting challenges and opportunities. We examine description-to-sketch system architectures and technologies such as CNNs, transformer-based text encoders, and GANs. A novel framework is proposed, combining description encoding, generative sketch rendering, and a feedback loop with facial recognition modules. Evaluated for sketch fidelity, identification accuracy, and computational efficiency, it addresses data scarcity and ethical issues, advancing practical forensic applications.

 


Full Text:

PDF

References


A. Shrivastava, T. Pfister, O. Tuzel, J. Susskind, W. Wang, and R. Webb, “Learning from simulated and unsupervised images through adversarial training,” Proc. IEEE Conf. Computer Vision and Pattern Recognition (CVPR), pp. 2107–2116, 2017.

Y. Zhang, P. Luo, C. C. Loy, and X. Tang, “From facial parts responses to face detection: A deep learning approach,” Proc. IEEE Int. Conf. Computer Vision (ICCV), pp. 3676–3684, 2015.

S. Zhang, Y. Zhu, C. Liu, Y. Chen, and S. Yi, “Deep face sketch synthesis via multi-task adversarial learning,” IEEE Trans. Image Processing, vol. 29, pp. 9541–9554, 2020.

J. Yu, C. Dong, J. Shi, and Y. Qiao, “Sketch-to-photo face synthesis using a conditional GAN,” Proc. ACM Int. Conf. Multimedia, pp. 1767–1775, 2019.

I. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, A. Courville, and Y. Bengio, “Generative adversarial nets,” Advances in Neural Information Processing Systems (NeurIPS), pp. 2672–2680, 2014.

H. Wu, H. Zheng, and H. Huang, “A deep learning framework for criminal identification from face sketches,” Pattern Recognition Letters, vol. 135, pp. 199–205, 2020.

Z. Wang, J. He, and X. Tang, “Learning structured representation for face sketch synthesis,” IEEE Conf. Computer Vision and Pattern Recognition (CVPR), pp. 4174–4182, 2019.

D. Yi, Z. Lei, and S. Z. Li, “Towards pose robust face recognition,” Proc. IEEE Conf. Computer Vision and Pattern Recognition (CVPR), pp. 3539–3545, 2013.

S. S. Rautaray and A. Agrawal, “Survey on deep learning in criminal investigation systems,” Big Data and Cognitive Computing, vol. 4, no. 3, pp. 1–19, 2020.

T. Karras, T. Aila, S. Laine, and J. Lehtinen, “Progressive growing of GANs for improved quality, stability, and variation,” Int. Conf. Learning Representations (ICLR), 2018.


Refbacks

  • There are currently no refbacks.