OpenCV Essentials: From Basics to Small Projects, by Irina Nikulina
Download as PPTX, PDF0 likes22 views
This presentation offers a practical introduction to OpenCV, one of the most widely used libraries for computer vision in Python. We'll explore essential tools and techniques useful for data scientists and engineers, and walk through two hands-on projects that demonstrate how OpenCV can be applied to real-world visual tasks.
OpenCV Essentials: From Basics to Small Projects, by Irina Nikulina
- 1. OpenCV Essentials: From Basics to Small Projects Irina Nikulina Paris 2025
- 2. My background E-commerce Creating Virtual Try-On solutions Dental Industry Digitizing crown production
- 3. Contents 1. OpenCV general functionalities 2. Image data pre-processing 3. Top-secret project #1 4. Classified project idea #2
- 4. OpenCV general info What is OpenCV? ❏ Open-source library ❏ Available for Python, C++, and other languages ❏ Includes several hundred of computer vision algorithms
- 5. OpenCV : Basic functions Easy image manipulation - great for simple visualisations
- 6. Images … and videos! ❏ Decompose a video into a sequence of frames ❏ Treat each frame as a separate image (NumPy array)
- 7. OpenCV for data processing - Geometric transformations - Color settings - Blurring / sharpening - Occluding / masking
- 8. OpenCV for data processing - Geometric transformations - Color settings (color spaces, brightness, etc.) - Blurring / sharpening - Occluding / masking
- 9. OpenCV for data processing - Geometric transformations - Color settings (color spaces, brightness, etc.) - Blurring / sharpening - Occluding / masking
- 10. OpenCV for data processing - Geometric transformations - Color settings (color spaces, brightness, etc.) - Blurring / sharpening - Occluding / masking
- 11. 3 reasons to do a side project ● To gain experience with new technologies ● To discover your passion ● To strengthen your CV
- 13. Panorama
- 14. Panorama Feature descriptors help describe image-specific regions around keypoints. They are invariant to geometric transformations and brightness variations.
- 15. Panorama Feature descriptors help describe a region so it can be found in other images
- 16. Panorama ? Matched descriptors help determine the geometric transformation between images.
- 17. Panorama Matched descriptors help determine the geometric transformation between images.
- 18. Panorama Matched descriptors help determine the geometric transformation between images.
- 19. Panorama
- 20. Panorama
- 21. Panorama Common framework for the image and 3D scans alignment Descriptors extraction Descriptors matching Data alignment Transformation estimation Applications: ❏ Robotics ❏ Medical Imaging ❏ 3D reconstruction
- 22. Panorama: ideas for improvement - Experiment with different feature descriptors to compare performance on various objects
- 23. - Experiment with different feature descriptors to compare performance on various objects - Enhance the final panorama: trim empty areas, adjust lightning and refine pivot image selection. Panorama: ideas for improvement
- 25. Face tracking OpenCV’s built-in object detection: - No GPU required - Easy to install and use - Works in real time (support video-streams) - Detects faces, eyes and smiles A great foundation for your project!
- 26. Smile tracking Raw detection without adjustments
- 27. Smile tracking - Tuned the parameters - Converted to grayscale - Applied image normalisation - Used face bounding box for smile detection
- 28. Smile tracking Raw detection without adjustments Detection with preprocessing
- 29. Face, eyes, smile tracking Put everything together:
- 30. Face, eyes, smile tracking Put everything together: Added a smile-triggered face-mask:
- 31. Blurring Face detection + blurring with Gaussian Blur
- 32. Blurring Face detection + blurring with Gaussian Blur
- 33. New features and ideas to level up your project: - Interactive face masks - Face-based authentication - Presence tracker, smile counter - Best photo selector - Emotion tracker / recognition (may require face landmark prediction from an external library) Be creative and enjoy the process!
- 34. New features and ideas to level up your project: - Interactive face masks - Face-based authentication - Presence tracker, smile counter - Best photo selector - Emotion tracker / recognition (may require face landmark prediction from an external library) Be creative and enjoy the process!
- 35. New features and ideas to level up your project: - Interactive face masks - Face-based authentication - Presence tracker, smile counter - Best photo selector - Emotion tracker / recognition (may require face landmark prediction from an external library) Be creative and enjoy the process!
- 36. New features and ideas to level up your project: - Interactive face masks - Face-based authentication - Presence tracker, smile counter - Best photo selector - Emotion tracker / recognition (may require face landmark prediction from an external library) Be creative and enjoy the process!
- 37. New features and ideas to level up your project: - Interactive face masks - Face-based authentication - Presence tracker, smile counter - Best photo selector - Emotion tracker / recognition (may require face landmark prediction from an external library) Be creative and enjoy the process!