Median Kernel Based Multi-Wavelet Feature Extraction and Multimodal Firefly Optimization and Learning Classification (MMFOLC) Knee Injury Detection

Abstract

Knee injuries are not easy to correctly diagnose. Knee injury has bad reaction on medical treatment and if the knee injury are not detected beginning stage, it will result in a growth in pain insecure. Increase cost relies on lots of subjective factors such as physician experience, swelling, patient guarding and the severity of the injury. Knee Injuries are the main cause of athlete’s players and can occur during the knee accident day to day activities. a physician manually manipulates the knee with a series of standard tests as well as knee injury increase the treatment cost. Our Proposed technique MMFOLC includes three type of process Median kernel Filter(MKF) pre-process and segment, Multi-wavelet feature extraction and Multimodal firefly optimization injury detection and classification. In Median kernel filter pre- processing technique decrease the noise image pixel and improve the image clearness quality. Second step move to the kernel image block segmentation it is used for reduce the error find knee injury detection in pixel based block segmentation. Next step Multi-wavelet feature extraction (MWFE) for detect knee injury unhealthy pixel detection shape and wavelet signal feature extraction its minimize the error and time consumption to detect the knee injury. Finally used Multimodal Firefly Optimization technique(MMFOLC) accurate Learning classify and detect the injury of Knee. this technique improve classification accuracy and reduce error rate compare with existing methods. Experimental evaluation of MMFO technique is carried out using a Knee injury data with different performance metrics such as accuracy, Prediction time, and true positive rate with respect to a number of Knee Osteoarthritis images