An effective mobile visual searching algorithm based on the bag-of-words method for furniture images

With the popularity of smart phones nowadays, image retrieval on mobile devices has been applied on a wider scale. To balance the efficiency and performance, an image retrieval algorithm which consists of two matching steps is proposed for furniture images captured by mobile devices (named mobile images). In the first stage, similar images are quickly obtained by the rough matching algorithm based on the bag-of-words (BOW) method. In the second stage, the most similar image is selected from the selection of similar images using the fine matching algorithm based on ORiented Brief (ORB).There are four main contributions in this paper. Firstly, add a spatial relationship to BOW features, an image is divided into three circular parts: the inner, middle and outer parts. Secondly, a feature point extraction method on concentric circles is proposed, which are resistant to geometric attack. Thirdly, the Gabor local line-based feature (GALIF) descriptor is improved by changing the filter object from the line to areas and overcome its weakness of low efficiency and rotation-sensitivity. Finally, a new image signature matching strategy is proposed based on Euclidean distance and a scale invariant feature to improve matching accuracy. Experimental results on a database with thousands of images show that our algorithm can retrieve mobile images quickly and precisely. © 2016.