Manuscript received April 24, 2025; revised June 4, 2025; accepted July 10, 2025; published September 17, 2025.

Abstract—Facial expression recognition is essential for a variety of applications, including marketing, commerce, security systems, and psychotherapy. The two primary categories of facial expressions are macro-expression and micro-expression. When comparing macro-expression, it can be difficult to recognize micro-expression in real time because it typically happens in high stakes scenarios. So, in this work, auto-detection of facial micro-expressions was proposed by using a Binomial Dropout Convolution Neural Network (BDCNN) classifier technique. Initially, the input video collected from the publicly available source is converted into frames and the keyframes are generated by using the Distance Deviation-based Rood Pattern Search (DDRPS) algorithm. Then, the keyframes are preprocessed to reduce the significance of illumination, face detection, landmark points on the face, and motion magnification. Afterward, the important features are extracted from the preprocessed frames. Finally, the BDCNN classifier was used to recognize the micro-expression. The average accuracy and F-Measure values obtained on the Spontaneous Actions and Micro-Movements (SAMM) dataset are 0.9758 and 0.9247; also, for the Spontaneous Micro-expression Corpus (SMIC) dataset, the values are 0.8989 and 0.8536, similarly, for the Chinese Academy of Sciences Micro-Expression (CASME) Ⅱ dataset, the accuracy and F-Measures are 0.9784 and 0.9509. The proposed classifier is compared to state-of-art methods to prove its superiority.

Keywords—micro expression, Distance Deviation based Rood Pattern Search (DDRPS) algorithm, Edge Preserving Homomorphic Filter (EPHF), YOLOv3, Linear Active Appearance Model (LAAM), Linear based Eulerian Video Magnification (LEVM), Binomial Dropout Convolutional Neural Network (BDCNN)

Cite: Anjani S. D. D, M. K. Kumar, Chinnam Sabitha, P. V. V. S. R. Kumari, and Sasi R. Desabathula, "Binomial Dropout Convolutional Neural Network Classifier-Based Micro-expression Recognition System," Journal of Image and Graphics, Vol. 13, No. 5, pp. 489-501, 2025.

Copyright © 2025 by the authors. This is an open access article distributed under the Creative Commons Attribution License (CC-BY-4.0), which permits use, distribution and reproduction in any medium, provided that the article is properly cited, the use is non-commercial and no modifications or adaptations are made.