Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{195624,
        author = {Nadimetla ShivaTeja and Jangalapelli Shiva and Deekonda Prem Kumar and Talari Pavan},
        title = {A Novel Framework Based on Deep Reinforcement Learning for Face Anti-Spoofing},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {11},
        pages = {1689-1694},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=195624},
        abstract = {There is a reduction in facial authentication due to emergence of other tricks of spoofing beginning with simple photo forgeries and progressing to very realistic deepfakes. CNNs are capable of performing a variety of tasks, but with some probability they create similar models and overlook novel types of fake images. I have observed that they gradually become ineffective as time passes. I created DRL-FAS to stop that. The former segment involves the reading of the overall image with a simple CNN. It determines the duration of standing, the place to search and the smallest details that are important. Such minor hints are fed into a RNN which learns through time. Lastly all these are compounded with general pointers in order to come up with a more powerful judgment. This will not stick to a given frame hence the model will continue to run throughout the video. We performed better in similar tests to those of Cai (in particular in against-untested attacks) in the group. The system identifies more deceptive fakes using less of the computing power.},
        keywords = {Face ID, machine learning, DRL-FAS, fraud (deception), trick of impersonation (spoof tricks).},
        month = {April},
        }