Face Expression Recognition based on Lucas-Kanade

Cheng Kong ； Xinyang Zhou ； Weiwei Cheng ； Zhe Sun

《International Core Journal of Engineering》 10卷2期 (2024/02) Pp. 212-218

https://doi.org/10.6919/ICJE.202402_10(2).0025

摘要

Aiming at the low recognition accuracy and efficiency of traditional convolutional neural network in Facial Expression Recognition (FER), a facial expression recognition method based on sparse optical flow method was proposed. First, the Haar feature extraction face recognition algorithm was used to detect and identify the target face in real time, and Harris corner detection was used to obtain the target face feature points to accurately locate the face; secondly, the optical flow tracking method was used to measure the movement trend of the target and calculate the target. The approximate position of the next moment, updated the feature points of the Harris corner algorithm and calculated the new tracking position to replace the original feature point data to ensure the accuracy of detection, and finally performd face position detection, normalization, and data enhancement on the image in the mini_Xception architecture. wait for preprocessing. The experimental results show that the recognition rate on the Fer2013 dataset is 68.50%.

關鍵字

Depthwise Separable Convolution ； Sparse Optical Flow Method ； Facial Expression Recognition

參考文獻

Chau M H, Jacobs G M. Applied Linguistics, language guidelines, and inclusive practices: The case for the use of who with nonhuman animals[J]. International Journalof Applied Linguistics, 2021, 31(2): 301-303.

Li S, Deng W. Deep facial expression recognition: A survey[J]. IEEE Transactions on Affective Computing, 2020.

Naidu P R, Sagar S P, Praveen K, et al. Stress Recognition Using Facial Landmarks and Cnn (Alexnet)[C]//Journal of Physics: Conference Series. IOP Publishing, 2021, 2089(1): 012039.

Zhou Jie, Ma Mingdong.Face Expression Recognition Based on Improved ResNet Network[J].Computer Technology and Development,2022,32(01):25-29.

Wu L, Lin X, Chen Z, et al. Surface crack detectionbased on image stitching and transfer learning with pretrained convolutional neural network[J]. Structural Control and Health Monitoring, 2021, 28(8): e2766.