Bat algorithm based Softcomputing Approach to Perceive Hairline Bone Fracture in Medical X-ray Images

Detection of bone fracture is easy if it is an open fracture and we do not need any special mechanism to identify that. But in case of closed fracture that is, hairline bone fracture the case becomes complicated. It is not always possible to detect that through bare eye. This is also a challenge for computer vision, through which detection of such cases could become easy and time saving. There are already a lot of methods and techniques available to solve this problem. But there is no such effort in the trend of using nature inspired metaheuristic algorithms. These emerging algorithms are becoming popular and can become a significant tool for computer vision. In this study, detection of hairline bone fracture is taken as a problem to be solved via implementing Bat algorithm. In the preprocessing stage this algorithm is been applied, to enhance the image, after which self organizing map (SOM) is used to draw segmentation and at last K-means clustering is used to produce the objective image. The results are exceptionally encouraging and the effectiveness of Bat algorithm as an image enhancer is proved. Keywords-medical imaging, hairline bone fracture, image processing, Bat algorithm, gamma correction, Self Organizing Maps, Kmeans, Peak Signal to Noise Ratio, histogram