Visual-based Items Recommendation Using Deep Neural Network

Online shopping websites throughout the world are increasing at an immense pace. Most of these websites are focused on search engines that tend to be based mainly on the information base and utilize matching keywords to find related products. However, consumers prefer a simple and secure digital set-up to search the related products. To address these problems, we propose a new solution to browse the items in an online buying system using a visual-based technique employing a deep learning approach. A consumer can supply or choose a visual, and related visual-based items will be presented to the consumer. The proposed approach encompasses two key steps: the first step is used for the classification and feature extraction. In the second step, the proposed recommendation system obtains similar matched related items. In the first step, the evaluation of the proposed model produces an 85.5% accurate model, thereby signifying its effectiveness for the items' recommendation in related applications.