Application of Clothing Design Based on Computer Vision Technology Using Hybrid Graph Convolution Neural Network and Lotus Effect Optimization Algorithm

The garment identification problem finds application in both computer vision and deep reading. In academic circles, clothing classification is a topic of great interest. This manuscript proposes an application of clothing design based on computer vision technology using hybrid graph convolution neural network (HGCNN) and lotus effect optimization algorithm (LEA) (ACD-CVTHGRNN-LEA). Initially, the extracted images from computer vision are collected from Fashion-MNIST dataset. Collected images are pre-processed to improve the quality of cloth design using hybrid graph transformer collaborative filtering (HGTCF). Later, preprocessed images are given to feature extraction; morphological features like shape, structure, colour, pattern, and size are extracted based on proportion-extracting synchrosqueezingchirplet transform (PESCT). Finally, the extracted features are fed to hybrid graph convolution neural network (HGCNN)for effectively classify the cloth design. In, general hybrid graph convolution neural network classifier does not express adapting optimization strategies to determine optimal parameters to ensure accurate cloth design detection system. Hence, the proposed method examined utilizing performance metrics like accuracy, precision, Fl-score, specificity, Recall, Least square error, and computation timing. Proposed ACD-CVT-HGRNN-LEA method attains 99% higher accuracy, and 150 (s) low times analysed to the existing methods respectively.