Advanced International Journal of Multidisciplinary Research

E-ISSN: 2584-0487

An Open Access Peer Reviewed Multidisciplinary Bi-monthly Scholarly International Journal

Call for Paper Volume 2 Issue 6 November-December 2024 Submit your research before last 3 days of December to publish your research paper in the issue of November-December.

Facial Expression Recognition with Convolutional Neural Networks

Author(s) Prakash Sangle
Country India
Abstract Facial expressions are one form of approach by which people convey their feelings, as they are a potent platform for interaction tool. Recognizing face expressions one of the exciting and effective jobs in public interaction since facial expressions are important in nonverbal interaction. Facial Expression Recognition (FER) is current study topic when it comes to AI, with numerous recent experiments utilizing CNN’s. In following study, it shows how to classify FER utilising CNNs and static pictures without performing any feature extraction or pre-processing work. The research also provides examples of pre-processing methods, such as face detection and lighting adjustment, to increase future accuracy in this field. The chin, mouth, eyes, nose, and eyebrows are among the most recognisable face characteristics that are extracted utilising feature extraction. We also talk about the literature review, our CNN design, the difficulties with max-pooling, and how dropout helped us get higher performance. In a classification job with seven classes, we achieved the efficiency of 61.7%in the FER2013 as opposed to the state-of-the-art classification accuracy of 75.2%.
Keywords Facial Action Coding System (FACS), Convolutional Neural Networks (CNN’s), Facial Expression Recognition (FER), Pre-processing, and Features of Extraction
Discipline Computer > Artificial Intelligence / Simulation / Virtual Reality
Published In Volume 1, Issue 1, July-August 2023
Published On 2023-07-10
Cite This Facial Expression Recognition with Convolutional Neural Networks - Prakash Sangle - AIJMR Volume 1, Issue 1, July-August 2023.

Share this