A Dynamic Programming Framework for Discovering Count and Values of Multilevel Image Thresholding

Researchers propose MET-DP, a dynamic-programming thresholding method that automatically infers how many thresholds an image needs.

The approach adapts minimum error thresholding to work across natural, satellite, and medical images without presetting threshold counts.

Experiments show MET-DP runs faster than traditional DP-based methods, especially at higher threshold counts.

However, some fixed-count methods still achieve better SSIM and PSNR, so the speedup comes with a quality tradeoff.