Everything you need to know about Adaptive Threshold
Adaptive Threshold converts your image to clean black and white by calculating a local threshold for every pixel. Unlike a simple global threshold, it automatically adapts to uneven lighting, shadows, and gradients — perfect for scanned documents, engravings, and high-contrast cutouts.
Parameters
- Radius
Controls the size of the local neighborhood analyzed for each pixel. Small values (1–3) preserve fine details. Larger values (8–12) smooth out lighting variations. - Bias
Shifts the threshold up or down. Negative values produce a darker result (more black). Positive values produce a brighter result (more white). - Smoothing
Softens the transition between black and white. Higher values reduce harsh speckling and jagged edges, creating smoother outlines. - Mix
Blends the black & white result with the original image. At 1.0, you see only the threshold effect. Lower values let the original colors show through.
Tips
- Great for extracting text from photos of documents or whiteboards
- Use a larger radius if lighting is very uneven across the image
- Add slight smoothing (0.01–0.05) to reduce noise in detailed areas