Gray-Level Image Transitions Driven by Tsallis Entropic Index
This work addresses image segmentation challenges for researchers in computer vision, but it appears incremental as it builds on existing maximum entropy principles.
The paper investigates how the Tsallis entropic index influences threshold values in image segmentation, finding that variations in the index can cause large leaps in thresholds, leading to abrupt transitions in gray-level images.
The maximum entropy principle is largely used in thresholding and segmentation of images. Among the several formulations of this principle, the most effectively applied is that based on Tsallis non-extensive entropy. Here, we discuss the role of its entropic index in determining the thresholds. When this index is spanning the interval (0,1), for some images, the values of thresholds can have large leaps. In this manner, we observe abrupt transitions in the appearance of corresponding bi-level or multi-level images. These gray-level image transitions are analogous to order or texture transitions observed in physical systems, transitions which are driven by the temperature or by other physical quantities.