🧠 Deep Learning Image Classification

🚀 Overview

This project implements an image classification system using Convolutional Neural Networks (CNN) with TensorFlow and Keras. The model is designed to automatically recognize and classify images into different categories through deep learning techniques.

It includes complete steps from data preprocessing to model training, evaluation, and prediction, making it a full end-to-end solution for image-based recognition tasks.

---

💡 Key Features

• 📁 Data Loading and Preprocessing

  • Handles image datasets with resizing, normalization, and augmentation.

• 🔍 Exploratory Data Analysis (EDA)

  • Visualizes image samples and class distributions for better dataset understanding.

• 🧠 Model Architecture (CNN)

  • Deep Convolutional Neural Network designed and trained using TensorFlow/Keras.
  • Uses Conv2D, MaxPooling, Dropout, and Dense layers for high performance.

• 🎯 Training and Evaluation

  • Tracks accuracy and loss during training.
  • Visualizes model performance on validation data.

• 🔮 Prediction and Visualization

  • Predicts image categories and displays prediction results with confidence scores.

---

⚙️ Technologies Used

• Python
• TensorFlow / Keras
• NumPy, Pandas, Matplotlib, Seaborn
• OpenCV (for image preprocessing)
• Scikit-learn (for evaluation metrics)

---

🧪 How It Works

1. Data Loading: Import image data and split into training and validation sets.
2. Preprocessing: Resize, normalize, and augment images to improve model generalization.
3. Model Building: Define and compile a CNN model using TensorFlow/Keras.
4. Training: Fit the model on training data and validate its accuracy.
5. Evaluation: Visualize metrics (accuracy, loss) and confusion matrix.
6. Prediction: Load new images and classify them with the trained model.

---

📈 Results

• High training and validation accuracy achieved using CNN.
• Model successfully generalizes to unseen images.
• Performance visualized using accuracy/loss plots and sample predictions.

Model and Presentation Links

Due to the large size of the ResNet50V2 model and presentation file, they are hosted on Google Drive. You can access them using the links below:

• ResNet50V2 Model: Download Here
• Presentation: Download Here