Authors:
Kapil Shrivastava, Eirtty Telang, Rahat Naz, Sony Kumari, Avdhesh Kumar Tiwari, Allen Paul Esteban

Addresses:
Department of Computer Engineering and Applications, GLA University Mathura, Mathura, Uttar Pradesh, India. Department of Computer Science and Engineering and Applications, Delhi Technical Campus, Gautam Budh Nagar, Uttar Pradesh, India. National School of Computer Science, University of Petroleum and Energy Studies, Dehradun, Uttarakhand, India. Department of Computer Science and Engineering -Artificial Intelligence and Machine Learning, ABES Engineering College, Ghaziabad, Uttar Pradesh, India. Department of Computer Science and Engineering, ABES Institute of Technology, Ghaziabad, Uttar Pradesh, India. Department of Graduate School Research Unit, Nueva Ecija University of Science and Technology, Cabanatuan, Nueva Ecija, Philippines.

Abstract:

Recent innovations in imaging technologies have brought about improved benefits for smart health care systems in enabling earlier detection and diagnosis of tumors and other diseases through imaging techniques. To develop effective treatment guidelines, tumor images must be accurately and precisely segmented and classified. To overcome limitations in current tumor image segmentation and classification algorithms regarding high levels of computational overhead, low levels of accuracy, and inability to classify complex tumor architectures correctly, this paper proposes enhancements to current algorithms for smart health care tumor classification. In particular, this paper discusses a tumor image segmentation algorithm that combines thresholding and region-growing methods, designed to reduce the computational overhead of current tumor segmentation algorithms while increasing the accuracy of tumor region identification. In addition, this paper presents a novel approach to feature extraction that focuses on the textural and geometric characteristics of tumor areas segmented. Compared with earlier algorithms for segmentation and classification, the RBF kernel has improved the accuracy of both segmentation and classification when used in conjunction with SVMs for tumor image recognition. Furthermore, the RBF kernel has increased the SVM's ability to recognise complex structures in tumour images. Consequently, the speed of determining whether a person has a tumour has increased, enabling patients to receive care earlier.

Keywords: Tumor Structures; Region-Growing; Segmentation Algorithm; Medical Imaging; Image Recognition; Feature Extraction; Textural Features; Computational Efficiency; Radial Basis Function (RBF).

Received on: 12/03/2025, Revised on: 15/05/2025, Accepted on: 07/08/2025, Published on: 08/03/2026

DOI: 10.69888/FTSHSL.2026.000593

FMDB Transactions on Sustainable Health Science Letters, 2026 Vol. 4 No. 1, Pages: 31-58

• Views : 96
• Downloads : 12

Download PDF