In this paper, we propose a solution for the problem of rotated partial shoeprint retrieval, based on the combined use of local points of interest and SIFT descriptor. Once the generated features are encoded using SIFT descriptor, matching is carried out using RANSAC to estimate a transformation model and establish the number of its inliers which is then multiplied by the sum of point-topoint Euclidean distances below a hard threshold. We demonstrate that such combination can overcome the issue of retrieval of partial prints in the presence of rotation and noise distortions. Conducted experiments have shown that the proposed solution achieves very good matching results and outperforms similar work in the literature both in terms of performance and complexity.
|Publication status||Published - 2009|
|Event||The 13th international Machine Vision and Image Processing Conference (IMVIP) - Dublin, Ireland|
Duration: 1 Jan 2009 → …
|Conference||The 13th international Machine Vision and Image Processing Conference (IMVIP)|
|Period||1/01/09 → …|