[NIR_face_dataset.zip] (NIR face dataset)
[gallery-groundtruth.txt] (gallery ground truth)
[probe-groundtruth.txt] (probe ground truth)
Dataset includes 3,940 NIR face images of 197 persons. The image size is 480 by 640 pixels, 8 bit, without compression.
The 3,940 images are divided into a gallery set and a probe set. In the gallery set, there are 8 images per person. In the probe set, 12 images per person.
The image information is provided in gallery-groundtruth.txt and probe-groundtruth.txt, which gives the image number, person number, and eye coordinates. Every line in the ground truth text file consists of the image file name and the eye positions. More specifically, every line in ground truth text is, [person id]-[image id for this person id]-[a][-g].bmp,[left eye x position],[left eye y position],[right eye x position],[right eye y position]. For example:
"0001-00-a-g.bmp,197,345,325,349" means: Person No.0001, Image No.00, with glasses; the left eye position is (197,345), and right eye (325,349)
"0007-22-a.bmp,175,182,285,179" means: Person No.0007, Image No.22, without glasses; the left eye position is (175,182), and right eye (285,179).
Return to Benchmark Collections