Plastic-Classification-Model

Project Overview

Plastic-Classification-Model is a deep learning project that classifies images of plastic waste as either Organic or Recyclable. The project uses a Convolutional Neural Network (CNN) built with TensorFlow/Keras, and provides a user-friendly web interface using Streamlit for real-time image classification.

---

What Was Done

1. Data Preparation & Exploration

• Dataset: Images of waste, categorized as Organic or Recyclable.
• Preprocessing: Images were resized, normalized, and augmented (rotation, zoom, flipping, etc.) to improve model generalization.
• Visualization: Data distribution and sample images were visualized to understand the dataset.

2. Model Building & Training

• CNN Architecture: Built a CNN with multiple Conv2D, MaxPooling, BatchNormalization, Dense, and Dropout layers.
• Training: Used data augmentation and split the data into training and validation sets.
• Optimization: Implemented EarlyStopping and ModelCheckpoint to avoid overfitting and save the best model.
• Evaluation: Trained the model for up to 30 epochs, monitored accuracy and loss, and visualized training history.

3. Model Evaluation

• Testing: Evaluated the model on a separate test set to check its generalization.
• Metrics: Used accuracy, classification report, and confusion matrix for performance analysis.
• Visualization: Displayed predictions and confusion matrix for better interpretability.

4. Deployment with Streamlit

• Web App: Developed a Streamlit app (app.py) that allows users to upload an image and get instant classification results.
• User Interface: The app displays the uploaded image, predicted class, and confidence score in a visually appealing format.

---

How to Access and Use This Project

1. Clone the Repository

git clone https://github.com/iarpitsaxena/Plastic-Classification-Model.git
cd Plastic-Classification-Model

2. Install Dependencies

Make sure you have Python 3.7+ installed. Install required packages:

pip install -r requirements.txt

3. Run the Streamlit App

Start the web application with:

streamlit run app.py

This will open a browser window with the Plastic Waste Classifier interface.

4. Use the Classifier

• Upload a clear image of plastic waste (jpg, png, jpeg).
• The app will display the image, predicted class (Organic or Recyclable), and the confidence score.

---

Functionality of Each Library/Package

• numpy: Efficient numerical computations and array operations, used for image data manipulation and preprocessing.
• pandas: Data analysis and manipulation, especially for handling labels and tabular data.
• matplotlib: Visualization of data distributions, training history, and prediction results.
• opencv-python: Image processing tasks such as reading, resizing, and color space conversion.
• tqdm: Progress bars for loops, especially during data loading and preprocessing.
• tensorflow: Deep learning framework used to build, train, and deploy the Convolutional Neural Network (CNN) model.
• streamlit: Framework for building and deploying the interactive web application for image classification.
• Pillow (PIL): Image loading and manipulation in the Streamlit app.

---

Files in the Repository

• plastic-waste-classification.ipynb: Jupyter notebook for data exploration, model training, and evaluation.
• app.py: Streamlit web app for image classification.
• best_model.keras / Waste-Classification-CNN-Model.h5: Saved trained models.
• requirements.txt: List of required Python packages.
• README.md: Project documentation.

---

Notes

• The model was trained on a dataset with two classes. For best results, use images similar to the training data.
• You can retrain the model with your own dataset by following the steps in the notebook.