Intellectual Property India Publishes Patent Application for Efficientnet-Based Deep Learning Framework For Automated Multi-Class Leukemia Detection And Classification Using Peripheral Blood Smear Images

MUMBAI, India, Aug. 17 -- Intellectual Property India has published a patent application (202641095535 A) filed by Dr. K Kalai Selvi; Sudha Krishnasamy; and Hariharan N on August 06, 2026, for Efficientnet-Based Deep Learning Framework For Automated Multi-Class Leukemia Detection And Classification Using Peripheral Blood Smear Images.

Inventors include Dr. K Kalai Selvi; Sudha Krishnasamy; and Hariharan N.

The application for the patent was published on August 14, 2026, under issue no. 33/2026.

Abstract: The present invention discloses an EfficientNet-based deep learning framework for the automated detection and multi-class classification of leukemia using peripheral blood smear images. The framework comprises an image acquisition module, an image preprocessing module, a data augmentation module, an EfficientNet-based feature extraction module, a transfer learning module, a multi-class classification module, a prediction module, and a clinical decision support module. The acquired blood smear images undergo preprocessing, including image resizing, normalization, standardization, and noise reduction, followed by data augmentation to improve model robustness and generalization. The preprocessed images are analyzed using pretrained EfficientNet architectures integrated with transfer learning to automatically extract discriminative morphological features associated with leukemia. A classification module incorporating Global Average Pooling, Batch Normalization, Dropout, and Softmax layers classifies the images into Acute Lymphoblastic Leukemia (ALL), Acute Myeloid Leukemia (AML), and Normal categories. The framework further generates diagnostic confidence scores and evaluates performance using standard classification metrics. The proposed invention improves diagnostic accuracy, reduces analysis time, minimizes observer variability, and provides a computationally efficient and clinically deployable computer-aided diagnostic solution for early leukemia screening, subtype classification, and treatment planning in hospitals, pathology laboratories, diagnostic centers, and telemedicine environments.

Disclaimer: Curated by HT Syndication.