Optimizing Depression Detection Through IVMPNet: A Fusion of Inception V4 and MLP

This research employs the CK+ (Cohn-Kanade Plus) dataset obtained from CMU's Perception Lab to enhance the detection of depression through the IVMP-Net framework, which integrates Inception V4 and Multilayer Perceptron (MLP) networks. A graphical user interface (GUI) developed in MATLAB streamlines the entire workflow, encompassing video input processing and the identification and categorization of facial emotions such as anger, disgust, happiness, neutrality, and sadness. By examining the prevalence of these emotional states, the system assesses depression levels and classifies them as mild, moderate, or severe. The methodology demonstrates an impressive overall accuracy of 98.9%, bolstered by precision, sensitivity, specificity, and F-score metrics that affirm the model's reliability. The analysis of the confusion matrix and performance indicators highlights the model's efficacy, suggesting its applicability for real-time emotion monitoring and automated mental health evaluations. This innovative approach provides a dependable mechanism for the early identification and intervention of depression, utilizing sophisticated deep learning methodologies to improve the precision and dependability of mental health assessments.