Sketch retrieval via local dense stroke features

Article history: Received 13 August 2014 Received in revised form 11 July 2015 Accepted 27 November 2015 Available online 22 January 2016 Sketch retrieval aims at retrieving the most similar sketches from a large database based on one hand-drawn query. Successful retrieval hinges on an effective representation of sketch images and an efficient searchmethod. In this paper, we propose a representation scheme which takes sketch strokes into account with local features, thereby facilitating efficient retrieval with codebooks. Stroke features are detected via densely sampled points on stroke lines with crucial corners as anchor points, from which local gradients are enhanced and described by a quantized histogramof gradients. A codebook is organized in a hierarchical vocabulary tree,whichmaintains structural information of visual words and enables efficient retrieval in sub-linear time. Experimental results on three data sets demonstrate the merits of the proposed algorithm for effective and efficient sketch retrieval. © 2016 Elsevier B.V. All rights reserved.