E-commerce Products Image Classification using EfficientNetB5 with Transfer Learning

The image classification is a research field where images are classified without human intervention. E-commerce image classification is the application of machine learning where techniques automatically categorize and label product images in E-commerce platforms. This study focuses on the complex task of e-commerce product image classification, which aims to develop a model capable of accurately classifying diverse products. This approach helps to take an acquired image from a designated device, decoding the acquired image, and applying a classification algorithm to categorize the products. The dataset has been organized into four categories, which are T-shirts, TVs, couches, and jeans. The E-commerce product dataset has been utilized with 796 images. The proposed work is based on transfer learning. In the proposed framework, pre-trained EfficientNetB5 model is used for fine-tuning for the e-commerce product classification task. The proposed framework achieved a 98% accuracy, 97.5% precision, 97% recall, and 97.75% F1- Score.