Intellectual Property India Publishes Patent Application for Ai-Enabled Adaptive Deep Learning Framework For Real-Time Skin Lesion Analysis And Early Skin Cancer Diagnosis

MUMBAI, India, July 24 -- Intellectual Property India has published a patent application (202641084992 A) filed by Dr. Ankur Kumar Meena; Mrs. R. Indupoornima; Ms. H. Sowmya; Ms. Sheeba D; Ms. P. Parameswari; Ms. D Betteena Sheryl Fernando; Dr. D. Justin Jose; Ms. Amudaria S; Ms. M Sugirtha Jasmine; Ms. K. Sahaya Sonia; Dr R N Devendra Kumar; and Mrs. Skb. Rathika on July 10, 2026, for Ai-Enabled Adaptive Deep Learning Framework For Real-Time Skin Lesion Analysis And Early Skin Cancer Diagnosis.

Inventors include Dr. Ankur Kumar Meena; Mrs. R. Indupoornima; Ms. H. Sowmya; Ms. Sheeba D; Ms. P. Parameswari; Ms. D Betteena Sheryl Fernando; Dr. D. Justin Jose; Ms. Amudaria S; Ms. M Sugirtha Jasmine; Ms. K. Sahaya Sonia; Dr R N Devendra Kumar; and Mrs. Skb. Rathika.

The application for the patent was published on July 17, 2026, under issue no. 29/2026.

Abstract: The present system discloses an AI-enabled adaptive deep learning framework for real-time skin lesion analysis and early skin cancer diagnosis. The framework integrates advanced image preprocessing, attention-guided lesion segmentation, hybrid handcrafted and deep feature extraction, adaptive feature fusion, multiclass deep learning classification, explainable artificial intelligence, continual learning, and cloud-assisted clinical decision support into a unified intelligent diagnostic system. Skin lesion images acquired from dermoscopic or clinical imaging devices are enhanced using contrast normalization, noise reduction, and artifact removal techniques before segmentation using attention-based neural networks. The segmented lesions are represented using complementary handcrafted and deep semantic features that are adaptively fused and classified using hybrid EfficientNet and Vision Transformer architectures. Explainable AI modules provide visual interpretation of diagnostic decisions, while adaptive learning mechanisms continuously improve model performance using validated clinical data without complete retraining. The framework further supports cloud-based teledermatology, mobile deployment, federated learning, and secure clinical integration, enabling accurate, interpretable, and scalable early skin cancer diagnosis suitable for real-time healthcare applications.

Disclaimer: Curated by HT Syndication.