deep learning apps
TheFoodINC – Smart Market Assistant
A deep learning web app that classifies grocery and market products from images and displays useful product information with confidence-based predictions.
PythonTensorFlowEfficientNetV2B0StreamlitDeep Learning
Model Summary
Project Overview
TheFoodINC is a deep learning web application that classifies grocery, fruit, and vegetable products from images. The goal is not only to predict the product class, but also to present the result in a user-friendly way with confidence scores, top-5 alternatives, and useful product information.
Project Goal
- Predict grocery and market products from uploaded images.
- Show model confidence clearly to the user.
- Display top-5 alternative predictions.
- Avoid overconfident results on unclear or out-of-dataset images.
- Provide useful product information cards with nutrition notes, storage tips, compatible foods, and recipe ideas.
Model Architecture
1.Input Image 224×224×3
2.Data Augmentation
3.EfficientNetV2B0 Backbone
4.GlobalAveragePooling2D
5.Dropout
6.Dense 63 Softmax
7.Top-5 Predictions