Pattern recognition and image segmentation based on some novel fuzzy similarity measures

Pattern recognition and image segmentation are challenging tasks in computer vision, with applications ranging from object detection to medical imaging. Fuzzy comparison measures emerge as an effective tool in handling these tasks, especially when two feature spaces are associated with uncertainty. Recently, many fuzzy comparison measures have been proposed and applied to different kinds of real-life problems. However, these measures have certain drawbacks because of their incapability to classify very similar pairs of objects, which may not produce effective results in the problems related to pattern recognition and image segmentation. This paper investigates the derivation of new fuzzy similarity measures. The proposed measures bear a continuous nature, allowing for subtle differences and smooth transitions between levels of similarity. This characteristic enhances the precise interpretation of linguistic variables. Moreover, pattern recognition and image segmentation techniques that primarily utilised proposed fuzzy similarity measures have been developed. Certain patterns and images are considered for the implementation of the proposed techniques. The results revealed that the proposed measures are advantageous over existing similarity measures, highlighting their potential for improving performance in pattern recognition and image segmentation.