This repository contains the implementation of a deep learning model for classifying different types of hybrid oranges. The project uses various advanced models like InceptionV3, MobileNet, and CNN for enhanced classification accuracy. The primary goal is to improve fruit quality control and optimize inventory management through automated classification.
- Project Overview
- Dataset
- Model Architecture
- Performance and Snapshots
- Installation
- Usage
- Contributing
- License
The project focuses on building a deep learning pipeline to classify hybrid oranges into different categories using image data. The pipeline involves data preprocessing, model training, evaluation, and performance tuning to achieve high accuracy.
- InceptionV3
- MobileNet
- Custom CNN
The dataset used for this project can be found on Kaggle. It contains images of various hybrid oranges, classified into different types.
Hybrid Oranges Dataset on Kaggle
To use the dataset in this project:
- Download the dataset from Kaggle.
- Extract the files and place them in the
data/directory of this repository.
The following models were implemented for the classification task:
- InceptionV3: Pretrained on ImageNet, this model was fine-tuned with hybrid orange data for improved accuracy.
- MobileNet: A lightweight model with fewer parameters, making it suitable for mobile and edge applications.
- Custom CNN: A custom-built convolutional neural network designed specifically for this classification task.
- InceptionV3: Achieved 95% accuracy on the validation set.
- MobileNet: Achieved 92% accuracy on the validation set.
- Custom CNN: Achieved 89% accuracy on the validation set.
To set up the project, clone the repository and install the required packages using requirements.txt.
Run the training script to start training the models and evaluate their performance.
Contributions are welcome! Feel free to open issues or submit pull requests.
This project is licensed under the MIT License. See the LICENSE file for more details.