About This Tool
Background Remover runs a browser-local foreground segmentation model, creates an alpha mask, and opens Advanced Cutout Studio for erase, restore, refinement, crop, background composition, and export.
What You Can Do
Automatic segmentation
Detect a likely foreground subject with the GXA Vision Model components.
Manual refinement
Erase, restore, refine, crop, and adjust the alpha mask.
Background composition
Keep transparency or apply a supported background before export.
How to Use
- 1
Choose a supported image for Background Remover
Choose a supported image for Background Remover.
- 2
Wait for the browser-local subject segmentation model to create the first mask
Wait for the browser-local subject segmentation model to create the first mask.
- 3
Use erase, restore, refinement, crop, and background controls in Advanced Cutout Studio as needed
Use erase, restore, refinement, crop, and background controls in Advanced Cutout Studio as needed.
- 4
Export the chosen transparent or composited result and inspect its edges at full resolution
Export the chosen transparent or composited result and inspect its edges at full resolution.
Practical Use Cases
Product cutouts
Separate a product from a plain backdrop for catalog layouts.
Portrait graphics
Prepare a subject for a profile design or presentation.
Privacy and File Processing
Background Remover keeps its JPG, JPEG, PNG, or WEBP input in browser memory while the operation runs. The workflow creates Transparent PNG, or Composited JPG/PNG where selected for local preview or download and does not send the entered values or selected file bytes to the GXA Toolbox application server.
Supported Inputs and Outputs
Input
- JPG
- JPEG
- PNG
- WEBP
Output
- Transparent PNG
- Composited JPG/PNG where selected
Quality and Fidelity
The segmentation mask is inferred from image content. Fine hair, translucent objects, motion blur, and low-contrast edges can require manual erase or restore refinement.
Limitations and Important Notes
- Maximum input is 30 MB and 48 megapixels.
- WebGPU availability varies; a WASM compatibility path is used when needed.
- Complex edges require visual review at full resolution.
Worked Example
Use a representative source image, apply Background Remover, and compare the result with the original at full size before publishing it.
Helpful Tips
- Keep the untouched source image while evaluating Background Remover.
- Check the Transparent PNG or Composited JPG/PNG where selected result at its intended display size.
Frequently Asked Questions
Does Background Remover overwrite the selected image?
No. Background Remover creates a result or extracted value without overwriting the selected source.
What affects Background Remover results?
The segmentation mask is inferred from image content. Fine hair, translucent objects, motion blur, and low-contrast edges can require manual erase or restore refinement.
Can a very large image fail in Background Remover?
Yes. Background Remover remains subject to browser decoding, canvas, model, and memory limits.