🔍 QualityLens

QualityLens is a computer vision application designed to classify fruit and evaluate its physical quality. By combining a custom-trained Convolutional Neural Network (CNN) with advanced image preprocessing, the app provides both species identification and a "Quality Score" percentage to help determine freshness and integrity.

💻 Technologies

• Python
• PyTorch (Custom CNN Architecture)
• FastAPI (Backend API)
• Streamlit (Frontend UI)
• OpenCV (Image Preprocessing)
• NumPy / Matplotlib (Data Analysis)

✨ Features

• Multi-Class Classification: Specifically trained to identify Apples, Oranges, and Bananas.
• Quality Scoring: Provides a percentage-based score (0-100%) evaluating the quality and freshness of the fruit.
• Custom Inference Engine: Uses a custom-built CNN architecture rather than a generic pre-trained model.
• Pre-processing Pipeline: Leverages OpenCV to normalize lighting and scale images for high-accuracy inference.

📁 The Process

To build QualityLens, I utilized the Kaggle Fruits Dataset to train a deep learning model from scratch. Unlike standard transfer learning, I designed a Custom Convolutional Neural Network (CNN) in PyTorch to understand specific spatial features of fruit textures, shapes, and surface defects.

[Image of a convolutional neural network architecture]

The workflow is split into three main stages:

1. Data & Training: Images were preprocessed using OpenCV for consistency (resizing, grayscale conversion, and normalization). The model was trained using PyTorch, optimizing for both classification accuracy and quality regression.
2. API Layer: A FastAPI server handles the model inference. It receives image data via POST requests, runs it through the saved .pth model weights, and returns a JSON response containing the label and the calculated quality percentage.
3. User Interface: A Streamlit dashboard allows users to upload images and visualize the results instantly with clean progress bars and confidence levels.

🚦 Running the Project

1. Clone the repository:

   git clone [https://github.com/Mickman7/QualityLens.git](https://github.com/Mickman7/QualityLens.git)

2. Install dependencies:

   pip install -r requirements.txt

3. Run the FastAPI Backend:

   uvicorn backend.main:app --reload

4. Run the Streamlit Frontend:

   streamlit run frontend/app.py

5. Analyze: Open your browser to the local Streamlit URL (usually http://localhost:8501) and upload a .jpg or .png of a fruit.

📺 Preview

(Replace this with a screenshot of your Streamlit dashboard showing the classification and quality score)

---