Authors:
Thirumalraj Karthikeyan, V. Revathi, Swathi Baswaraju, S. Venkatasubramanian, Kawsher Rahman

Addresses:
Quest Technologies Research Labs, Quest Technologies, Trichy, Tamil Nadu, India. Department of Research and Development, New Horizon College of Engineering, Bengaluru, Karnataka, India. Department of Computer Science and Engineering, New Horizon College of Engineering, Bengaluru, Karnataka, India. Department of Computer Science and Business Systems, Saranathan College of Engineering, Trichy, Tamil Nadu, India. Department of General Medicine, Beanibazar Cancer and General Hospital, Sylhet, Bangladesh.

Abstract:

Women often die from ovarian cancer (OC). Recent studies show deep learning can better predict OC phases and subtypes. This study predicts OC phases using a free Cancer Genome Atlas dataset (TCGA-OV) website. Pre-processing uses Mean-Median-Gaussian (MMG), a new hybrid filtering approach. The predictive model is strengthened and more accurate by combining three basic filtering approaches. Key data is collected via Faster SqueezeNet. This feature speeds up the extraction of information from complex genetic data. Ovarian cancer stages are then accurately classified using the EfficientNet-V2 network. This improves diagnosis. Understanding the importance of patient data security, this research proposes a simple, lightweight image encryption method. The Lorenz Chaotic System and DNA coding are combined. This concealment strategy protects patient data during the investigation. There's also optimal key selection using the Child Drawing Development Optimisation Algorithm. Better encryption keys boost safety. Ovarian cancer detection is complete with these approaches. It diagnoses accurately and protects patient data. The Proposed EfficientNet model achieves 98.88% accuracy in ovarian cancer classification, outperforming other models in terms of precision, recall, specificity, and F1 score. The suggested Lightweight Encryption model offers high security, an entropy of 6.9, a PSNR of 40.5, and efficient encryption and decryption.

Keywords: Ovarian Cancer; Mean Median; Gaussian Filter; Squeeze Network; Efficient Net; DNA Coding; Child Drawing; Development Optimisation; Lorenz Chaotic System, Hybrid Filtering.

Received on: 03/09/2024, Revised on: 20/11/2024, Accepted on: 04/01/2025, Published on: 05/06/2025

DOI: 10.69888/FTSHSL.2025.000459

FMDB Transactions on Sustainable Health Science Letters, 2025 Vol. 3 No. 2, Pages: 65-79

• Views : 39
• Downloads : 10

Download PDF