Brad Ulery, R. A. Hicklin, Craig I. Watson, Michael D. Indovina, Kayee K. Hanaoka
The Slap Fingerprint Segmentation Evaluation 2004 (Slap Seg04) was conducted to assess the accuracy of algorithms used to segment slap fingerprint images into individual fingerprint images. Segmenters from ten different organizations were evaluated on data from seven government sources, according to several distinct measures of accuracy. The source of data, the segmentation software used, and the decision criteria used were each found to have a significant impact on accuracy. Depending on the data source, the best segmenters produced at least 3 matchable fingers, with finger positions correctly identified, from 93% to over 99% of the slaps. The source of data is a much better predictor of success than whether the images were collected on livescan devices or paper. Most segmenters performed well, but there were significant differences among segmenters on poor quality data.