Therapeutic Effect of Visual Expressive Arts Based on Deep Learning

Abstract

Art therapy is an innovative and versatile form of psychotherapy that utilizes the creative process of art making as a therapeutic tool to address emotional, psychological, and social challenges. This approach acknowledges the inherent connection between art and human expression, dating back to ancient times when art was used for healing and self-discovery. In this manuscript, Therapeutic effect of visual expressive arts based on deep learning (TE-VEA-DHNN) is proposed. Initially, the images are collected from WikiArt Emotions dataset are given as input. The input images are fed to pre-processing using Confidence partitioning sampling filtering (CPSF) for remove the background noise from the input image. Afterward the pre-processed image is given to Synchro Transient Extracting Transform (STET) for extracting the texture features such as entropy, contrast, correlation and Homogeneity. Then the extracted features are given to Dense Hebbian Neural Network for predicting the artistic expressions and provide therapists with insights into their emotional states. In general,18:05:24 Dense Hebbian Neural Network (DHNN) does not express adapting optimization strategies to determine optimal parameters to ensure accurate prediction based on visual expressive arts therapy. Hence, the Archerfish Hunting Optimizer (AHO) is to optimize to DHNN which accurately predict the visual expressive arts therapy. The proposed TE-VEA-DHNN approach is implemented in Python. The performance of the proposed TE-VEA-DHNN approach attains 23.26%, 24.37% and 25.97% higher accuracy and 21.73%, 23.84% and 25.87% higher recall compared with existing methods such as application of deep learning in art therapy (AT-CNN), An art therapy evaluation method based on emotion recognition using EEG deep temporal features(ATEM-LSTM) and A Portrait of Emotion: Empowering Self-Expression through AI-Generated Art (ESE-AI-GA) respectively.