🐶🐱 Image Classification of Dogs vs Cats using KNN and HOG Features

This project performs image classification on the popular Dogs vs Cats dataset using a K-Nearest Neighbors (KNN) classifier. Feature extraction is done using Histogram of Oriented Gradients (HOG), and visualization is performed using confusion matrices and classification reports.

---

📁 Contents

• Vaishnavi_imageClassification.ipynb – Notebook implementing:
  • Dataset download via Kaggle
  • Image preprocessing
  • HOG feature extraction
  • KNN training and evaluation

---

📌 Workflow Summary

1. Download Dataset

   • Automatically downloads dogs-vs-cats dataset from Kaggle using API token.

2. Image Preprocessing

   • Resizes images to 64x64
   • Loads and batches ~300 sample images

3. Feature Extraction

   • Converts images to grayscale
   • Extracts HOG features using skimage.feature.hog

4. Classification

   • Splits data (80% train / 20% test)
   • Trains a KNN classifier (with k=3)
   • Evaluates performance

5. Visualization

   • Classification report
   • Confusion matrix heatmap

---

🛠️ Dependencies

Install the following packages before running the notebook:

pip install numpy opencv-python scikit-learn scikit-image tensorflow matplotlib seaborn

---

📂 Dataset Details

• Dataset: Dogs vs Cats
• Structure:

  dataset/
  └── dogs_vs_cats/
      ├── train/
      │   ├── cat.0.jpg
      │   ├── dog.0.jpg
      ├── test/
      │   ├── cat.1.jpg
      │   ├── dog.1.jpg
  

---

🚀 How to Use (in Google Colab)

1. Upload your kaggle.json token (from your Kaggle account API).
2. Run the notebook cells in order.
3. Outputs:
   • HOG feature-based KNN model performance
   • Confusion matrix heatmap
   • Precision, recall, and F1-score

---

📊 Results

• Classifier: K-Nearest Neighbors (k=3)
• Feature Extractor: HOG
• Evaluation:
  • Confusion Matrix
  • Classification Report

---