By Topic

Gait Recognition Using Shadow Analysis

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$31 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

2 Author(s)
Iwashita, Y. ; Grad. Sch. of Inf. Sci. & Electr. Eng., Kyushu Univ., Fukuoka, Japan ; Stoica, A.

The exploitation of biometrics information in human shadow silhouettes (shadow biometrics), derived from video imagery after processing by gait analysis methods opens new avenues in remote biometrics. For the first time it becomes possible to obtain "overhead" biometrics, which may lead to recognition of human identity and behavior from high altitude airborne platforms using overhead video sequences. "Shadow biometrics" may use shadow information without body information, or in combination with it, as an additional perspective approximately equivalent to the use of a second camera. We took recordings and created a gait database in which both shadows and bodies are visible and used it to provide a first demonstration of the human gait recognition from shadow analysis. Only the information from shadows was used, which appears appropriate for overhead surveillance. We select a set of horizontal on the silhouette, for which we determine their length; this determines a set of varaibles in time, to which we apply spherical harmonics for each gait cycle. A k-nearest neighbor classification is applied to spherical harmonic coefficients. A subset of only 5 different subjects were used in this work to avoid biasing the results since we did not compensate for changing of sun position with time; the correct classification rate (CCR) was 95%. In additional tests, we reduce spatial and temporal resolution of the images to 50% each, which reduced the CCR to 75%.

Published in:

Bio-inspired Learning and Intelligent Systems for Security, 2009. BLISS '09. Symposium on

Date of Conference:

20-21 Aug. 2009