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{194483,
        author = {Janhvi Hiwe and Dipti Mehare and Rutuja Jogi and Anjali Tamte and Vaishnavi Mahulkar and Sakshi Likhitkar and Sneha Deshmukh},
        title = {A Multi-Stage Deep Learning Framework for Face Detection and Facial Landmark Localization Using MTCNN},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {10},
        pages = {5353-5358},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=194483},
        abstract = {Detecting human faces in images is an important task in the field of computer vision. It is widely applied in systems such as security surveillance, biometric verification, attendance tracking, and human–computer interaction. However, traditional face detection approaches often face difficulties when images contain different lighting conditions, head poses, image scales, or partially hidden faces. To address these challenges, deep learning techniques have been introduced. One such approach is the Multi-task Cascaded Convolutional Neural Network (MTCNN), which is designed to detect faces while also identifying important facial landmarks. This paper studies the structure and working of the MTCNN model in detail. The architecture is composed of three sequential neural networks: the Proposal Network (P-Net), the Refinement Network (R-Net), and the Output Network (O-Net). Each stage progressively improves the accuracy of face detection and removes incorrect detections. The performance of the model is analyzed using well-known datasets such as the WIDER FACE dataset and the FDDB dataset. In addition, the effectiveness of MTCNN is examined by comparing it with other popular models like YOLO, FaceNet, and Dlib. The findings show that the cascaded architecture of MTCNN helps achieve reliable face detection accuracy while maintaining reasonable computational efficiency. Overall, the study demonstrates that MTCNN is a practical solution for face detection tasks in real-world scenarios},
        keywords = {Face Detection, MTCNN, Deep Learning, Computer Vision, Convolutional Neural Networks.},
        month = {March},
        }