Analyzing fashion trends using CNN and Cloth classification with respect to season and state

Machine learning now does have a broad aspect of apps in a variety of fields. Deep learning methods, particularly convolution neural networks (CNN), are frequently used to analyse visual imagery. CNN main applications include face recognition, image recognition, object recognition, and so on. As the use of machine learning in many industries grows rapidly around the world, the fashion industry is following suit. The proposed model describes the creation of a computer vision system for clothing detection and classification in e-commerce images. The architectures used in this work for detection and classification are YOLO v3 and Residual Networks, respectively. We are training and testing the clothes detection and classification networks using a portion of the Fashion image dataset, which contains box annotations for clothing location, as well as manually collected data. The proposed models detect clothing using bounding boxes. The dataset consists of 8 different classes with 500 images per class. Investigational outcomes show that, as anticipated, CNN algorithms are the most effective and efficient network configurations for clothing detection and classification. Following testing, it was discovered to have 90% accuracy. In addition, Indian, Western, and Indo-Western cuisines are classified by state and season.