Graph matching for object recognition and recovery

A robust skeleton-based graph matching method for object recognition and recovery applications is presented. The object model uses both a skeleton model and contour segment models, for object recognition and recovery. Initially, the skeleton representation is created from the input contour that is provided by a deformable contour method (DCM). The skeleton is then matched against a set of object skeleton models to select a candidate model. Corresponding feature points or landmarks on the input and model contours are determined from their skeletons automatically. Based on the landmarks, the input and model contours are broken into contour segments. The input contour segments are then matched against the corresponding model segments for error analysis. For any large error in the segment mismatch, a fine-tuning process is performed to enhance the final segmentation result. The skeleton-based shape matching algorithm is illustrated by using a set of animal silhouette examples. Experiments of object recovery using real biomedical image samples, such as MRI knee and brain’s corpus callosum images, have shown satisfac-