Image inpainting is the process of removing damage, such as noises, strokes or text, on images. It is particularly useful in the restoration of old photographs which might have scratched edges or ink spots on them. These can be digitally removed through this method.
Image inpainting works by replacing the damaged pixels with pixels similar to the neighboring ones, therefore, making them inconspicuous and helping them blend well with the background. Consider the image below.
The image has some marks to the right. To inpaint this image, we require a mask, which is essentially a black image with white marks on it to indicate the regions which need to be corrected. In this case, the mask is created manually on GIMP.
Inpainting Algorithms –
OpenCV implements two inpainting algorithms:
- “An Image Inpainting Technique Based on the Fast Marching Method”, Alexandru Telea, 2004:
This is based on Fast Marching Method (FMM). Looking at the region to be inpainted, the algorithm first starts with the boundary pixels and then goes to the pixels inside the boundary. It replaces each pixel to be inpainted with a weighted sum of the pixels in the background, with more weight given to nearer pixels and boundary pixels.
- “Navier-Stokes, Fluid Dynamics, and Image and Video Inpainting”, Bertalmio, Marcelo, Andrea L. Bertozzi, and Guillermo Sapiro, 2001:
This algorithm is inspired by partial differential equations. Starting from the edges (known regions) towards the unknown regions, it propagates isophote lines (lines that join same-intensity points). Finally, variance in an area is minimized to fill colors.
FMM can be invoked by using
cv2.INPAINT_TELEA, while Navier-Stokes can be invoked using
cv2.INPAINT_NS. The Python code below inpaints the image of the cat using Navier-Stokes.
Attention geek! Strengthen your foundations with the Python Programming Foundation Course and learn the basics.
To begin with, your interview preparations Enhance your Data Structures concepts with the Python DS Course. And to begin with your Machine Learning Journey, join the Machine Learning – Basic Level Course