Efficacious retrieval of similar shapes from large image databases is still a challenging problem. In recent works about shape retrieval, methods based on Dynamic Space Warping (DSW) and descriptors with contour information have had a significant presence. This paper introduces a technique for Content-Based Image Retrieval (CBIR) that encompasses a robust corner detector and a new shape descriptor based on local and global features invariant to translation, rotation and scale. Moreover, this technique employs the effective DSW tool for shape matching and retrieval. We have conducted our experiments on binary images from MPEG-7 and Tari 1000 databases and our experiments have shown that the proposed method performed well in comparison with other methods defined by salience points and triangle-area representation.