Object &
Watermark Remover

What Is Inpainting, and How Does It Remove Objects Without Touching the Rest of the Photo?

Inpainting is a computer-vision technique that reconstructs a masked region of an image by analysing the texture, colour and structure of the pixels surrounding it. Rather than cropping, blurring or covering the area, the algorithm essentially asks "what would likely be here based on everything around it?" and fills the gap accordingly — which is exactly how professional photo editors' "content-aware fill" and "spot healing" tools work under the hood.

This tool uses the Telea inpainting algorithm from OpenCV, running entirely inside your browser via WebAssembly — the same underlying library used in countless professional computer-vision applications. You paint a mask over the region you want gone, and the algorithm works outward from the edges of that mask, propagating nearby colour and texture information inward until the area is filled. Because only the masked pixels are touched, every other pixel in your photo remains byte-for-byte identical to the original.

Results are best on simple, repetitive or low-detail backgrounds — skies, walls, grass, water, fabric. Highly detailed or structured backgrounds (faces, text, fine patterns) are inherently harder for any inpainting algorithm to reconstruct convincingly, so for tricky spots it often helps to remove smaller sections at a time using the "Touch Up Further" option, rather than painting one large area in a single pass.