Step 1: Upload Image
Click to upload image
JPG, PNG, WebP up to 10MB
How it Works:
- 1. Upload an image with objects to remove
- 2. Use the brush to paint the areas you want removed
- 3. Use the eraser to fix any mistakes
- 4. Click "Use This Mask" when done
- 5. Adjust settings and click "Remove Object"
- 6. Download your cleaned image
Features
- • Telea and Navier-Stokes inpainting methods
- • Multiple removing methods
- • Adjustable parameters
- • Real-time preview
- • Mask-guided processing
Supported Formats
Common browser-supported image formats
How to remove an object from an image
Upload an image, paint over the area you want to remove, and confirm the mask. Choose an inpainting method and radius, then start processing. The image and mask are sent to the server, where OpenCV fills the painted region using surrounding image information.
How the painted mask works
The mask tells the processor exactly which pixels should be replaced. White painted areas mark regions to remove, while black areas are kept. The mask is generated at the original image dimensions before it is sent for processing.
Paint slightly beyond the object edge
Cover the complete object and a small amount of its boundary. The backend expands and smooths the mask before inpainting, which helps avoid leaving thin fragments of the selected object around its edges.
Telea vs Navier-Stokes inpainting
Telea
Telea is the default method. It fills the masked region by propagating nearby image information inward from the mask boundary. It is generally a practical first choice for relatively small unwanted areas.
Navier-Stokes
Navier-Stokes inpainting uses a different OpenCV method for continuing nearby image structure into the selected region. If Telea leaves an obvious artifact, compare the Navier-Stokes result on the same mask.
What the inpaint radius controls
The radius determines how far around each masked pixel OpenCV looks for neighboring image information. The current tool accepts values from 1 to 20 pixels. A small radius may work well around fine details, while a larger radius uses a wider neighborhood. Bigger values do not automatically produce a better result.
Best suited to smaller removal areas
Classical inpainting works best when the masked region is reasonably small and surrounded by useful nearby texture. Small blemishes, wires, marks, isolated objects, and simple background interruptions are generally easier than large complex subjects.
Large or complex objects can be difficult
The processor does not invent a new scene or understand hidden background content. Large masks, faces, detailed architecture, repeated patterns, or objects covering important structures may produce visible smearing or reconstructed textures that do not match the original scene.
How the mask is prepared before removal
Before inpainting, the server converts the mask to a binary image, performs morphological closing and opening, expands the selected region, smooths mask edges, and thresholds it again. This preprocessing helps create a cleaner removal boundary before the selected inpainting algorithm runs.
Tips for better object removal
Output and processing details
The image and mask are processed on the server. The current backend returns the finished result as a JPEG image. Temporary input, mask, and output files used by the API are deleted after processing completes.
Remove Object FAQ
Does the tool automatically detect the object?
No. You manually paint a mask over the area you want removed, giving you direct control over which part of the image is processed.
Is this AI object removal?
The current implementation uses classical OpenCV inpainting rather than an object-detection or generative AI model.
Which inpainting method should I choose?
Telea is a good first choice. If the result is not satisfactory, compare Navier-Stokes using the same mask and radius.
What radius should I use?
Start with a moderate value. The radius controls the neighboring area used during inpainting, and values from 1 to 20 pixels are supported.
Can it remove a very large object?
It may process the mask, but large removal areas are more likely to show artifacts because classical inpainting depends on nearby visible image information.
What format is the result?
The current server route returns the processed image as JPEG.