Overview
This repository contains the code and models for our submission to the Flipkart Grid competition. Our solution focuses on two key features:
Object Detection: Accurately identifies and locates objects within images, including fruits, vegetables, and other grocery items. Object Classification: Categorizes detected objects into specific classes (e.g., apple, banana, tomato) and brands. Object Counting: Determines the quantity of each object category present in the image.
Freshness Assessment: Evaluates the freshness level of fruits and vegetables using image analysis techniques. Variety Handling: Accommodates different types of fruits and vegetables, considering factors like color, texture, and shape. Model Architecture
Our solution leverages a combination of state-of-the-art machine learning and deep learning techniques: YOLOv8: A powerful object detection model for accurate and efficient object localization and classification. Transfer Learning: Utilizes pre-trained models on large datasets to improve performance and reduce training time. Convolutional Neural Networks (CNNs): Extracts relevant features from images, such as color, texture, and shape. Transfer Learning: Leverages pre-trained models like ResNet or EfficientNet for feature extraction. Custom Classifier: Trains a classifier to predict freshness levels based on extracted features. Streamlit: A user-friendly Python library for building web applications.