28-04-2017, 01:57 PM
29-04-2017, 11:23 AM
Bone fracture is a common problem in humans occurs due to the high pressure applied in the bone or simple accident and also due to osteoporosis and bone cancer. Therefore the exact diagnosis of bone fracture is important aspects in medical field. In this work, X-ray / CT images are used for the analysis of bone fractures. The objective of this project is to develop an efficient system based on image processing for a fast and accurate classification of bone fractures based on information obtained from X-ray / CT images.
Images of fractured bone are obtained from the hospital and processing techniques such as pre-processing, segmentation, edge detection and feature extraction methods are adopted. The processed images will also be classified into fractured and non-fractured bone and will compare the accuracy of the different methods. This project is completely used MATLAB 7.8.0 as the programming tool for image upload, image processing and user interface development. The results obtained demonstrate the performance of the bone fracture detection system with some limitations and a good accuracy of 85%.
X-ray images are examined manually, but they are time-consuming and prone to errors. As X-ray images are more suspicious of noise we have used many pre-processing steps to eliminate noise and blur of the image. Therefore, the system is able to detect the fracture with greater precision. The system detects fracture based on the type of fracture. The noise is removed from the image and the image is transformed into a clearer image so that the system can easily detect the fracture. We use the image processing methodology to trace the bone. All unwanted and small objects are deleted by the system.
Finally based on the connected component, the system detects the fracture. The system shows the bounding box around the fracture. This system involves steps of image preprocessing and fracture detection based on the type of fracture. The proposed system is able to detect bone dislocation with a success rate of 80% and greater bone fracture 60-70% accuracy and minor fracture with 50-60% accuracy.
Images of fractured bone are obtained from the hospital and processing techniques such as pre-processing, segmentation, edge detection and feature extraction methods are adopted. The processed images will also be classified into fractured and non-fractured bone and will compare the accuracy of the different methods. This project is completely used MATLAB 7.8.0 as the programming tool for image upload, image processing and user interface development. The results obtained demonstrate the performance of the bone fracture detection system with some limitations and a good accuracy of 85%.
X-ray images are examined manually, but they are time-consuming and prone to errors. As X-ray images are more suspicious of noise we have used many pre-processing steps to eliminate noise and blur of the image. Therefore, the system is able to detect the fracture with greater precision. The system detects fracture based on the type of fracture. The noise is removed from the image and the image is transformed into a clearer image so that the system can easily detect the fracture. We use the image processing methodology to trace the bone. All unwanted and small objects are deleted by the system.
Finally based on the connected component, the system detects the fracture. The system shows the bounding box around the fracture. This system involves steps of image preprocessing and fracture detection based on the type of fracture. The proposed system is able to detect bone dislocation with a success rate of 80% and greater bone fracture 60-70% accuracy and minor fracture with 50-60% accuracy.
It can be understood in the following video: