A Web Based Fruit Classification Platform with Deep Learning Approach
Loading...
Date
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
UNIVERSITI MALAYSIA SARAWAK
Abstract
Description
Accurate identification of fruit species is made extremely difficult by their global range, especially when it comes to unfamiliar or uncommon kinds. Traditional fruit categorization techniques frequently result in inefficiency and human mistake since they mainly rely on manual observation and experience. To overcome these challenges, a web-based fruit categorization platform that makes use of deep learning technologies is suggested. Through the ability to input fruit images for immediate identification, the proposed solution streamlines the procedure and produces precise and effective outcomes. An agile design process was used to create the system, enabling continuous feedback and incremental improvements while it was being developed. To improve classification performance, a pre-trained EfficientNetB2 model was modified using the Fruits-100 dataset. To further enhance the user experience, the platform also includes a structured database with details fruit information. The final implementation demonstrates high responsiveness and accuracy. Other features, such as the ability to create PDF reports and track history, significantly improve usability. Future enhancements might include expanding the fruit dataset and adding ripeness detection capabilities. By bridging the gap between traditional classification methods and modern AI-driven solutions, this project develops a fruit classification platform that is scalable, effective, and user-friendly.