System for sickle cell disease classification and severity assessment using YOLOv8, has image preprocessing pipeline involving normalization, segmentation, and feature extraction, classification module utilizing YOLOv8 to differentiate between normal and sickled red blood cell based on morphology
	SARASWATHI T; RAGUL S; PRABAKARAN M
	2025-04-01
专利权人	EASWARI ENG COLLEGE (EASW-Non-standard)
申请日期	2025-04-01
专利号	IN202541032198-A
成果简介	NOVELTY - System for sickle cell disease classification and severity assessment using YOLOv8 comprises an artificial intelligence (Al)-powered deep learning model trained on blood smear image datasets to detect and classify red blood cells (RBCs). An image preprocessing pipeline involving normalization, segmentation, and feature extraction to enhance detection accuracy. A classification module utilizing YOLOv8 to differentiate between normal and sickled RBCs based on morphological characteristics. A severity assessment module calculating the percentage of sickled cells in a given sample and categorizing disease progression into normal, mild, moderate, and severe levels. An Al-powered recommendation system, integrated with the Google Gemini API, to generate personalized treatment suggestions based on the detected RBC abnormalities. USE - System for sickle cell disease classification and severity assessment using YOLOv8. ADVANTAGE - The system categorizes sickle cell disease (SCD) into normal, mild, moderate, and severe stages, aiding clinical decision-making, integration with Google Gemini API provides Al powered recommendations for personalized treatment. The performance evaluations demonstrate 98 % classification accuracy and 95 % efficiency, outperforming traditional methods. A Stream lit-based interface ensures accessibility. The future enhancements focus on improving accuracy, expanding datasets, and ensuring compliance with healthcare standards. The study advances AI-driven hematological diagnostics, transforming SCD detection and personalized patient care. The system using YOLOv8-based classification model processes segmented RBCs applies deep learning techniques to accurately differentiate between normal and sickled cells. The severity assessment module quantifies the proportion of sickled cells in a blood smear and classifies the disease into different severity levels, aiding in clinical decision-making. The Al-powered recommendations are generated based on the severity assessment, utilizing machine learning models to provide personalized treatment suggestions for patients. DETAILED DESCRIPTION - An INDEPENDENT CLAIM is included for the method for preprocessing of blood smear images involves undergoing contrast enhancement, normalization, and segmentation techniques, and improving image compact discs (CD) quality and feature extraction.
IPC 分类号	C07K-014/805 ; G01N-001/28 ; G01N-015/14 ; G16H-050/20 ; G16H-070/60
国家	印度
专业领域	信息技术
语种	英语
成果类型	专利
文献类型	科技成果
条目标识符	http://119.78.100.226:8889/handle/3KE4DYBR/13253
专题	中国科学院新疆生态与地理研究所
作者单位	EASWARI ENG COLLEGE (EASW-Non-standard)
推荐引用方式 GB/T 7714	SARASWATHI T,RAGUL S,PRABAKARAN M. System for sickle cell disease classification and severity assessment using YOLOv8, has image preprocessing pipeline involving normalization, segmentation, and feature extraction, classification module utilizing YOLOv8 to differentiate between normal and sickled red blood cell based on morphology. IN202541032198-A[P]. 2025.

条目包含的文件
条目无相关文件。

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。