Browse Prior Art Database

Method and System for Detecting Bone Fracture in X-ray images through Random Forest Fusion Technique

IP.com Disclosure Number: IPCOM000241432D
Publication Date: 2015-Apr-25
Document File: 4 page(s) / 127K

Publishing Venue

The IP.com Prior Art Database

Abstract

A method and system is disclosed for detecting bone fracture in X-ray images through random forests fusion technique.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 01 of 4

Method and System for Detecting Bone Fracture in X-ray images through Random Forest Fusion Technique

Musculoskeletal system, in medical domain, is defined as the combination of the muscular and skeletal systems. The musculoskeletal system provides our bodies with shape, protection of internal organs and the ability to move. Due to the dominancy of musculoskeletal system over the human body, the prevalence of musculoskeletal medical condition ranks at the top on the list, as compared with other categories of disease. Among different kinds of musculoskeletal injuries, the top two are: 1) muscle strain; and 2) bone fracture. Even ranked as the second, the bone fracture clearly requires more medical attentions compared with muscle strain. In clinical practice, the detection of bone fracture is often through visual inspection of X-ray images by radiologists to determine the presence and severity of the fracture, and the required course of treatment. However, bone fractures are also sometimes missed or misdiagnosed during the radiological examination due to human error and increases during the course of a long day at the radiology reading room. So there is need for a method and system that allows radiologists to detect bone fractures more efficiently and accurately.

Disclosed is a method and system for detecting bone fracture in X-ray images through random forests feature fusion technique. The method and system provides a generalized bone fracture detection technique that is applicable to multiple bone fracture types and multiple bone structures throughout the body. The detection technique intakes various types of training patches with different features to train random forests classifiers. The various types of training patches

with different features can be a positive (fracture) or a negative (healthy) image. Thereafter, the trained random forests classifier produces confidence score maps for local patches within the image, which indicate the probability of a patch

containing a fracture. Subsequently, based on the confidence scores map, Efficient Subwindow Search...