Copyright © 2025 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{177473,
        author = {Bhuvanesh P and Jansi Rani S and Hari Krishna P A and Dinesh Kumar B},
        title = {OPTIMIZING USER RECOGNITION WITH DEEP CNN AND CROPPED FACIAL INPUTS},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {11},
        number = {12},
        pages = {751-755},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=177473},
        abstract = {Systems for facial recognition have become widely used in fields including attendance tracking, security, and authentication. However, different lighting conditions, background noise, and facial occlusions make it difficult to achieve great accuracy in real-world circumstances. To improve recognition accuracy and model efficiency, this research proposes a deep learning-based method that makes use of a Convolutional Neural Network (CNN) trained just on cropped face areas. Prior to feature extraction, faces are cropped to improve attention on discriminative facial traits and remove irrelevant background data. After separating and aligning face areas using a face detection pipeline, the system sends the information to a deep CNN that has been trained on a labelled dataset for classification. Results from experiments show that, especially in hardware-constrained contexts, the suggested approach greatly enhances recognition performance when compared to models trained on full-frame images. This study offers insights into creating lightweight yet precise recognition algorithms and emphasizes the significance of region-focused preprocessing in deep facial recognition.},
        keywords = {CNN, Cropped Face Images, Face Detection, Haar Cascade, Recognition},
        month = {May},
        }