Loading...
Please wait, while we are loading the content...
Similar Documents
Amalgamation of features for latent fingerprint segmentation
| Content Provider | Indraprastha Institute of Information Technology, Delhi |
|---|---|
| Author | Vashisth, Tarun Jain, Aayush |
| Abstract | Latent fingerprints, found at crime scenes, typically suffer from background noise thereby making task of any automatic latent fingerprint identification system difficult. In order to increase the efficiency of automatic latent fingerprint identification system, region segmentation is performed which involves extracting the boundary of ridge information from a given image. However, manual segmentation of latent fingerprints is not feasible for large scale matching applications and automated approaches are not efficient against challenges arising due to background noise in latent fingerprint images. In this paper, we introduce a novel machine learning algorithm for automatic segmentation of latent fingerprints. An image is tessellated into local patches and a set of features are extracted, based on which a region/patch is classified as background or foreground using Random Decision Forest classifier. The results on three publicly available databases suggest that the proposed algorithm is able to efficiently segment even in the presence of various types of irregularity in latent fingerprint images and outperforms existing latent fingerprint segmentation algorithms. |
| File Format | |
| Language | English |
| Access Restriction | Authorized |
| Subject Keyword | Latent Fingerprint Image Analysis Machine Learning Saliency Random Decision Forests |
| Content Type | Text |
| Educational Degree | Bachelor of Technology (B.Tech.) |
| Resource Type | Thesis |
| Subject | Data processing & computer science |
