Threshold Segmentation
Segment foreground from background with global and adaptive thresholding for document cleanup, mask generation, region separation, and binarization workflows.
Threshold segmentation is useful when the key task is separating dark from light regions, isolating foreground objects, or preparing a binary image for later morphology, OCR, contour analysis, or connected components. It is one of the most common preprocessing steps in computer vision.
How to Use Threshold Segmentation
- Upload an image: Upload one of the formats listed by the tool (JPG, PNG, WebP) and wait for the local preview.
- Adjust settings and process: Choose the thresholding method (Binary, Adaptive, Otsu, Niblack, Sauvola, Wolf, or NICK) and adjust the parameters.
- Review and download: Review the result, then download it in one of the formats offered by this tool.
Frequently asked questions
What is image thresholding?
Thresholding converts a grayscale image to binary (black and white) by comparing each pixel to a threshold value. Pixels above the threshold become white, below become black.
What is Otsu thresholding?
Otsu's method automatically determines the optimal threshold by minimizing the intra-class variance between foreground and background. It works best when the histogram has two distinct peaks (bimodal distribution).
When should I use adaptive thresholding?
Adaptive thresholding is best for images with uneven lighting, shadows, or varying backgrounds. It calculates a different threshold for each local region of the image.
What are Niblack, Sauvola, Wolf, and NICK?
These are advanced local binarization methods from OpenCV ximgproc. They compute per-pixel thresholds using local mean and standard deviation. Sauvola and Wolf are generally better for document images. NICK works well with low-contrast text.
Is the Threshold Segmentation tool free?
Yes. All tools on Pixlane are free with no signup required. It runs entirely in your browser.
What does the block size parameter control in adaptive thresholding?
Block size defines the local neighborhood window used to compute the threshold for each pixel. Larger blocks smooth out local variations (better for noisy images); smaller blocks adapt to finer detail (better for complex documents with varying text sizes).