Transparent Peer Review By Scholar9
Facial Expression Recognition Using Deep Learning Algorithm
Abstract
Facial emotion recognition extracts the human emotions from the images. As such, it requires an algorithm to understand and model the relationships between faces and facial expressions, and to recognize human emotions. Recently, deep learning models are extensively utilized enhance the facial emotion recognition rate. However, the deep learning models suffer from the overfitting issue. Moreover, deep learning models perform poorly for images which have poor visibility and noise. Therefore, in this dissertation, a novel deep learning based facial emotion recognition tool is proposed. Initially, a joint trilateral filter is applied to the obtained dataset to remove the noise. Thereafter, contrast-limited adaptive histogram equalization (CLAHE) is applied to the filtered images to improve the visibility of images. Finally, a deep convolutional neural network is trained. Nadam optimizer is also utilized to optimize the cost function of deep convolutional neural networks. Experiments are achieved by using the benchmark dataset and competitive human emotion recognition models. Comparative analysis demonstrates that the proposed facial emotion recognition model performs considerably better compared to the competitive models.
Chandrasekhara (Samba) Mokkapati Reviewer
24 Sep 2024 05:40 PM
Approved
Relevance and Originality
This dissertation addresses a significant area in artificial intelligence by focusing on facial emotion recognition, which has broad applications in fields like psychology, human-computer interaction, and security. The integration of advanced techniques such as trilateral filtering and CLAHE demonstrates an original approach to enhancing model performance, particularly in challenging conditions like noise and low visibility.
Methodology
The methodology should be clearly articulated, detailing each step of the proposed tool’s development, including the dataset used, preprocessing techniques (trilateral filtering and CLAHE), and the architecture of the deep convolutional neural network (CNN). Additionally, it would be beneficial to specify the parameters and configuration of the Nadam optimizer to provide insight into the training process and optimization strategies employed.
Validity & Reliability
To support the validity of the findings, the dissertation should discuss the evaluation metrics used to assess model performance, such as accuracy, precision, recall, and F1 score. Including comparisons with established models will enhance the credibility of the results. Moreover, addressing potential limitations related to the dataset, such as size and diversity, will help contextualize the findings and their applicability in real-world scenarios.
Clarity and Structure
The paper should be organized into distinct sections, including an introduction, methodology, results, and discussion, to facilitate comprehension. Each section should transition smoothly, guiding readers through the research process and findings. Ensuring clarity in language and avoiding excessive technical jargon will make the dissertation accessible to a broader audience, enhancing engagement with the topic.
Result Analysis
The result analysis should provide a comprehensive examination of the experimental outcomes, highlighting the improvements achieved through the proposed methods compared to existing models. Including visual representations of results, such as confusion matrices or performance graphs, would enrich the analysis. Discussing the implications of these findings for future research and practical applications in emotion recognition systems will further underscore the relevance of the work.
IJ Publication Publisher
Done Sir
Chandrasekhara (Samba) Mokkapati Reviewer