Object detection in X-ray images based on object candidate extraction and support vector machine

In this paper, an adaptive region-growing method based on image histogram is used in image segmentation. The seeds are selected according to image histogram distribution. After segmentation, object candidates are obtained by combining the segmented regions according to the proposed rules. Finally, the object detection is completed by using Support Vector Machine. The proposed methods of image segmentation and object candidates extraction are simple and easy to operate, and the candidate extraction needs less requirement of segmentation. The exact shape of the candidate can be gained rather than the rectangle of the object. The experimental results prove that the correct results are obtained by using the proposed methods.