| Abstrakt: |
Biomedical image segmentation is considered an important and challenging task. Automated biomedical image analysis plays a major role in the early and quick diagnosis of diseases. Accurate and precise segmentation can lead to early treatment planning and it demands sophisticated approaches. Inspired by this, a novel approach is proposed. This approach will be known as the Fuzzy modified cuckoo search with spatial exploration (FMCSSE). High correlation among pixels is an important property of image data and pixels surrounding a particular pixel possess similar feature information. Therefore, it is extremely essential to consider the spatial information to generate a meaningful segmented image. The traditional fuzzy clustering approach is not suitable for exploiting spatial information. Therefore, this work is designed to explore spatial information and find the optimal clusters from biomedical images with the help of the fuzzy-modified cuckoo search approach. This approach is applied to different biomedical images and compared with various state-of-the-art unsupervised approaches like FEMO, FMCS, MCS, and CS. The proposed approach does not suffer from the choice of the initial assignment of the cluster centers. The proposed approach uses the type-2 fuzzy system blended with the modified cuckoo search (McCulloch approach) and spatial exploration procedure. Both qualitative and quantitative results show the superiority of the FMCSSE approach in terms of performance. [ABSTRACT FROM AUTHOR] |
Full text from SpringerLink