Remove Image Background by Color: what this tool does
Sample a background from the image corners or preview, make similar colors transparent, and refine the edge with tolerance and softness. The source is decoded and processed in the current browser tab; CalculatorGeek does not receive the image bytes.
This page owns the remove image background by color task. For the adjacent workflow, use Photo Editor, or return to the Image Tools Hub to choose another operation.
For the next related calculation, use the Convert JPG to PNG or WebP. Create PNG, WebP or browser-supported AVIF copies of JPG files, optionally limiting their maximum dimensions.
How to use Remove Image Background by Color
- Choose a supported source image or enter the page-specific source content.
- Set the visible controls for this task.
- Inspect the processed preview, dimensions, format and measured byte size.
- Run the final action and download the result. Keep the original until you have checked the output.
To compare this result with another relevant method, open the Watermark Image. Apply a text or logo watermark to one image or a batch with position, rotation, size, opacity and tiling controls.
Method and result meaning
The tool can average four corner pixels, sample the top-left pixel or use a color picked from the original preview. Every pixel is compared with that RGB value; tolerance removes close colors and softness creates partial alpha at the edge.
The result panel is the authoritative state for this session: it reports the rendered dimensions, encoded format and actual output bytes. The source file is never overwritten.
Which settings should you choose?
Choose settings from the destination requirement, not from a promise of universal quality.
- Start with Average all four corners for a studio-style background that reaches every corner.
- Use Pick from original preview when the background color is elsewhere or the corners differ.
- Increase tolerance gradually; use softness to reduce a harsh cutout edge without removing more interior colors.
Worked example
For a product photo on a nearly white background, select #ffffff and begin with tolerance 36. Increase it only until the remaining off-white edge disappears, checking that light areas inside the product remain intact.
After completing this task, Crop Image handles the next distinct image operation without mixing its settings into this page.
Privacy, supported files and limitations
This is color-based removal, not automatic subject segmentation. Busy or similarly colored backgrounds require a dedicated masking editor.
Processing is local to the browser. Each input is limited to 25 MB and 80 megapixels, each output to 64 megapixels, and batch-capable tools accept up to 20 files. Animated inputs use the decoded still frame.
Inputs, outputs and result meaning
The tool can average four corner pixels, sample the top-left pixel or use a color picked from the original preview. Every pixel is compared with that RGB value; tolerance removes close colors and softness creates partial alpha at the edge.
Inputs
- Source width
- Decoded source width in pixels. Supported units are px.
- Source height
- Decoded source height in pixels. Supported units are px.
Outputs
- Decoded source pixels
- This is the primary result. It is displayed in px. The display rule uses up to 0 decimal places.
The documented method is applied after the inputs are validated and normalized. Keep the result label, unit, selected mode, and stated limitations together when sharing the answer.
Frequently asked questions
Does the tool detect the subject automatically?
No. It samples a background color and applies a visible color-distance rule; it does not claim AI subject segmentation.
Which output format keeps transparency?
The tool exports PNG because JPG cannot store transparent pixels.
Sources and method transparency
The implementation and its stated limits were reviewed on October 14, 2026 against the following browser-platform documentation.
- MDN - CanvasRenderingContext2D.drawImage() - Documents the browser drawing, scaling and source-rectangle operation used by the image pipeline.
- MDN - HTMLCanvasElement.toBlob() - Documents browser canvas encoding and format fallback behavior.
- MDN - CanvasRenderingContext2D.getImageData() - Documents reading pixel color and alpha values for color-key transparency.