Dive into the captivating realm of computer vision and artificial intelligence with our Image Classification from Video project! 🎥🤖 This project is more than just recognizing faces; it's an adventure into the heart of cutting-edge technologies. From extracting frames to training models, this project is your ticket to the future of computer vision.
- Video Processing: Swiftly download and process videos using Python scripts.
- Face Detection: Harness the power of Haar Cascade Algorithm to extract faces from images and videos.
- Deep Learning Magic: Enchanting pre-trained FaceNet models and the mystical VGGFace architecture for precise face recognition.
- Modular Mastery: Explore organized modular code for a seamless understanding and effortless customization experience.
- Data Visualization: Peer into the world of embeddings and model predictions with mesmerizing visualizations.
- Google Colab Integration: Experience the magic of Google Colab for efficient and cloud-powered model training.
Our dataset is a collection of frames from the beloved sitcom show, Friends. Characters like Rachel, Chandler, Phoebe, Monica, and Ross are the stars of this dataset. With a total of 35 images (7 per person for training and 15 for testing), this dataset forms the bedrock of our image classification odyssey.
- Language: Python (Version 3.6.2)
- Libraries: OpenCV, scikit-learn, numpy, os, pytube, scikit_image, skimage, keras, tensorflow
- OpenCV: Your trusty sidekick for image and video processing.
- Haar Cascade Algorithm: The mystical spell for accurate face extraction.
- Pre-trained FaceNet Model: Unleashing the power of embeddings for face recognition.
- VGGFace Architecture: Deep learning sorcery for the most precise identification spells.
- Google Colab: Your enchanted notebook for cloud-powered, robust models.
- input: The treasure chest containing training and validation data, downloaded videos, and extracted frames.
- src: The spellbook, housing the core modular codebase for video processing, face extraction, and model training.
- output: The magical repository for predicted frames and model outputs.
- prebuilt_models: The ancient scrolls containing pre-trained models and your very own enchanted, trained models.
- lib: The magical library filled with reference Jupyter notebooks.
- requirements.txt: The potion recipe listing all the magical ingredients. Use
pip install -r requirements.txtto concoct your magic potion!
- Clone the Repository:
git clone https://github.com/your-username/image-classification-from-video.git- Install Dependencies:
pip install -r requirements.txt- Explore the Jupyter Notebooks: Dive into the src folder and explore the modular Jupyter notebooks for detailed insights.
For questions, collaborations, or further information, feel free to reach out:
Project Reference: Project Pro
LinkedIn: Vidhi Waghela
Email: [email protected]