Pants Style Classification Using YOLOv8: An Approach for Enhanced Detection of Pants

In recent years, the fashion industry has witnessed a significant shift towards automation and digitization, particularly in garment classification and detection. This study aims to implement YOLOv8 to detect six types of pants styles: straight jeans, skinny jeans, slacks, jogger pants, cargo pants, and overalls. A portable device was created using the Raspberry Pi Camera Module 3 to capture live feeds and the Raspberry Pi 4 to classify pants styles. Datasets sourced from online repositories were employed and trained using the Python programming language with the TensorFlow library. A comprehensive testing phase, comprising three hundred samples with fifty samples per classification, was conducted. Evaluation using a confusion matrix showed the system achieved an accuracy rate of 88%. These findings highlight the potential of YOLOv8-based pants style classification in various applications, including automated inventory management, personalized shopping experiences, and fashion trend analysis.