Enabling Social Interaction: A Face Recognition System for Visually Impaired People using OpenCV

  • Adedayo A. Sobowale Dept. of Computer Engineering,Fed. Poly Ede, Osun state
  • T. A. Abdul-Hameed
  • Peace O. Sobowale
  • Bolaji Johnson

Abstract

Over the last few decades, there have been substantial developments in a variety of domains, including computer science, artificial intelligence, and machine learning, which has accelerated the evolution of intelligent systems. Examples include speech recognition system, face recognition. This research work developed a method to assist blind and visually impaired people in the aforementioned forms of social interactions in order to address all these deficiencies.  This device can recognize faces of individuals by silently broadcasting their names over speakers using face recognition technology. The system uses a camera to capture an image of a person's face, which is then processed to extract key features such as the eyes, nose, and mouth. The extracted features were compared to a database of known faces to identify the person in the image. Blind and visually impaired individuals encounter significant challenges in identifying people during social interactions. Traditional methods like speech recognition might not be reliable in all situations, such as with silent group members. This social isolation can hinder their participation in professional and educational settings. This research proposes a novel face recognition system to address these challenges. The system utilizes a camera to capture a person's face. Key facial features are then extracted using the Open Computer Vision Library (OpenCV). These features are compared against a pre-enrolled database of known faces for identification. Upon successful recognition, the system discreetly announces the person's name through audio output. This system empowers visually impaired individuals to navigate social interactions more confidently. By providing a reliable method for facial recognition, the system promotes greater social inclusion and participation in various environments. The model shows excellent performance, consistently achieving high accuracy, peaking at 89.1% on the 69th epoch, and consistently maintaining high validation accuracy, reaching 91.2% in the 67th epoch, indicating its ability to function effectively.

References

Ajiroba, A., Lee, J., Billinghurst, M., Jhajharia, S., Pal, S. K., & Verma, S. (n.d.). (2019) Wearable Computing and its Application.

Bai, J., Lian, S., Liu, Z., Wang, K., & Liu, D. (2020). Combines deep learning-based face recognition with wearable technology, emphasizing practical implementation.. https://arxiv.org/abs/2005.02305

Bengio, Y.,Goodfellow, I., & Courville, A. (2016). Deep learning. MIT press.

Bledsoe, W. W. (1964). Semiautomatic Facsimile Transmission of Pictures. Proceedings of the Eastern Joint Computer Conference, 225-230.

George, A., & Ravindran, A. (2022). Focuses on integrating face recognition into smart glasses, using algorithms optimized for speed and efficiency on low-power devices. https://ieeexplore.ieee.org/document/9807113

Jain, A., Learned-Miller, E., & Li, H. (2016). Twin faces revisited. IEEE Transactions on Pattern Analysis and Machine Intelligence, 38(10), 2027-2039. doi: 10.1109/TPAMI.2015.2512825.

Kim, J., Lee, H., Kim, D., & Kim, J. (2018). Facial Recognition-Based Navigation System for the Visually Impaired. Proceedings of the IEEE International Conference on Computer Vision (ICCV), 2018, 123-130.

Karthigaikumar, P.,Francis, G. A., & Kumar, G. A. (2020). Face recognition system for visually impaired people. International Journal of Innovative Technology and Exploring Engineering, 9(2), 1090-1094.

LeCun, Y., Jarret, K., Kavukcuogle, K., & Ranzato, M. (2015). What is the best multistage architecture for object recognition? Proc.International conference on Computer Vision (ICCV’15).

Liu, W., Wen, Y., Yu, Z., Li, M., Raj, B., & Song, L. (2017). SphereFace: Deep hypersphere embedding for face recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (pp. 212-220).

Russell, S. J., & Norvig, P. (2018). Artificial intelligence: A modern approach. Pearson Education.

Sarwar, S., Qaisar, S. M., & Pasha, M. (2023). Explores the use of deep learning (ResNet50) for both face recognition and expression analysis, providing the visually impaired with richer social cues. https://www.researchgate.net/publication/373895875_Face_and_Facial_Expressions_Recognition_System_for_Blind_People_Using_ResNet50_Architecture_and_CNN

Shi, X., Liu, Y., Zhao, Y., Zhang, M., & Huang, Q. (2019). A Wearable Assistive Device for Face Recognition in Individuals with Visual Impairments. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 27(4), 636-643.

Turk, M., & Pentland, A. (1991). Eigenfaces for Recognition. Journal of Cognitive Neuroscience, 3(1), 71–86.

Zhang, H., Li, Y., Wang, X., & Liu, Q. (2016). A Multitask Approach for Simultaneous Face Detection and Alignment using Cascaded Convolutional Networks. Journal of Computer Vision and Pattern Recognition, 45(7), 987- 1003.

Published
2024-07-01