Sugeno integral generalization applied to improve adaptive image binarization

Adaptive binarization methodologies threshold the intensity of pixels with respect to adjacent exploiting integral images. In turn, images are generally computed optimally using summed-area-table algorithm (SAT). This document presents a new adaptive technique based on fuzzy through an efficient design modified SAT for integrals. We define this methodology as FLAT (Fuzzy Local Thresholding). The experimental results show that proposed have produced image quality thresholding often better than traditional algorithms and saliency neural networks. propose generalization Sugeno CF 1,2 integrals improve existing computation. Therefore, these generalized can be used tool grayscale processing in real-time deep-learning applications. Index Terms: Image Thresholding, Processing, Fuzzy Integrals, Aggregation Functions