Publication Date

Spring 2023

Degree Type

Master's Project

Degree Name

Master of Science (MS)


Computer Science

First Advisor

Nada Attar

Second Advisor

Wendy Lee

Third Advisor

William Anderopoulos


Convolutional Neural Network, Data Augmentation, Deep Learning, Facial expression recognition, Image Processing, Normalization


Facial expression recognition (FER) has been a challenging task in computer vision for decades. With recent advancements in deep learning, convolutional neural networks (CNNs) have shown promising results in this field. However, the accuracy of FER using CNNs heavily relies on the quality of the input images and the size of the dataset. Moreover, even in pictures of the same person with the same expression, brightness, backdrop, and stance might change. These variations are emphasized when comparing pictures of individuals with varying ethnic backgrounds and facial features, which makes it challenging for deep-learning models to classify. In this paper, we provide a simple yet efficient way for recognizing facial expressions that combines a CNN with certain image pre-processing techniques. We conducted our experiments on a combination of MUG, JAFFE, and CK+ datasets. To improve the performance of CNN, we experimented with various image pre-processing techniques such as face detection and cropping, image sharpening using Unsharp Mask, and normalization techniques like Global Contrast Normalization, Histogram Equalization, and Adaptive Histogram Equalization. Furthermore, we also examined data augmentation techniques such as image translations and adding noise to images to enhance performance of the deep learning model. Our custom CNN-based FER model achieved a maximum average accuracy of 93.3% (6 classes) and 91% (7 classes) after cross-validation. Our experimental results show that our proposed method can effectively enhance the accuracy of facial expression recognition