Classification of Muscle Contusion Healing Stages using Ultrasound Statistical Parameters and Image Texture Features Extraction

Abstract

Contusion is a common injury corresponding to rapid and strong compressive force during exercise occasions The contusion may be diagnosed clinically via the evaluation of medical images that however require the diagnosis by experienced and professional experts In the present study efforts were made to extensively assess the recovery of muscle tissues which were brought out following a certain contusion hit by a custom-made contusion model via quantitative parameters and texture feature of the acquired ultrasound radio-frequency (RF) signals and images The experiments were performed from lateral gastrocnemius muscle of SD rats in which the contusion was achieved by a hitting of 600g mass from a height of 30cm The integrated backscatter (IB) statistical Nakagami parameter and texture features extraction using gray-level co-occurrence matrix (GLCM) were calculated from the acquired ultrasound signals and B-mode images with a 30 MHz central frequency Finally support vector machine (SVM) classifier was used to classify the contusion recovery stages through quantitative parameters and texture features from ultrasound images The change of m and IB values which can clearly differentiate between destruction stages and other stages (p

Date of Award	2018 Aug 30
Original language	English
Supervisor	Shyh-Hau Wang (Supervisor)