The Intelligent Investigating K-Nearest Neighbors for Image Segmentation

Jamsandekar P.P.; Mishra S.; Hemelatha S.; Parida P.K.; Docente E.A.A.N.; Docente E.J.H.F. · Proceedings - 2023 IEEE International Conference on Paradigm Shift in Information Technologies with Innovative Applications in Global Scenario, ICPSITIAGS 2023 (2023) · DOI: 10.1109/icpsitiags59213.2023.10527676

Resumen

The sensible Investigating ok-Nearest buddies for photograph Segmentation (MIKKNS) is a brand new method for the automated photograph segmentation task. It combines the traditional and ok-nearest friends (KNN) set of rules with contemporary intelligent algorithms, which include deep learning, to deal with the challenges of guide segmentation in scientific imaging. It combines the two to enhance the segmentation accuracy even with fewer schooling parameters. Similarly, MIKKNS introduces a singular studying framework that uses a convolutional neural network (CNN) for characteristic extraction and the KNN to produce an optimized segmentation. Empirical opinions reveal that the proposed method outperforms traditional strategies by up to two. 3% in segmentation overall performance.

Instituciones: Universidad Autónoma de Ica