1 / 12

Comparison Results of Visual and Thermal face images Using FaceIt 4/02/03

Comparison Results of Visual and Thermal face images Using FaceIt 4/02/03. Jingu Heo The department of Electrical & Computer Engineering The University of Tennessee, Knoxville. Ranks and Confidence Rates. Ranks : the candidates that best match an unknown person presented to the database.

moor
Download Presentation

Comparison Results of Visual and Thermal face images Using FaceIt 4/02/03

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Comparison Results of Visual and Thermal face images Using FaceIt 4/02/03 Jingu Heo The department of Electrical & Computer Engineering The University of Tennessee, Knoxville

  2. Ranks and Confidence Rates Ranks : the candidates that best match an unknown person presented to the database Confidence Rates : Shows how confident the results are ( 10 : Same Image) Equal Error Rate : 8 ( False Acceptance Rate= False Rejection Rate)

  3. Confidence Rates – How 2 images are similar The performance of face recognition system can be varied by the choice of gallery and probe images.

  4. Database from Equinox Co-registered Visual and Thermal images • Total DB ( 1,643*2 = 3,206) Eye Aligning • Video : Automatic • Thermal : Manaul

  5. Visual and Thermal images withoutGlasses Probe 2 Probe 1 Gallery

  6. Comparison Results (Ranks and Confidence Rate) Thermal images has higher confidence rates and converges within 10 matches.

  7. Comparison Results(Ranks) – Glass off

  8. Comparison Results(Ranks) – Glass on Visual : Glasses slight affect the performance Thermal : Glasses significantly affect the performance The confidence rate is lower than 7.0 : Cannot believe the results) Center of the glasses = center of eye positions?.

  9. Visual and Thermal images with Glasses Low Confidence High Confidence

  10. Automatic Eye detection in the thermal images Challenging problem -Even if we detect eyes, the performance will be poor with current FaceIt Algorithms • Co-registered Visual and Thermal images (Same Eye Coordinates) • Gallery : Without glass (Visual +Thermal) • Probe : Fusion based on Confidence Rate • Glasses on : Recognition using Visual • Glasses off : Recognition using Thermal or fused results • of visual and thermal confidence.

  11. Advantage/Disadvantages

  12. Future Works • Registration of Visual and Thermal image • Hardware only: Co-registration (Visual camera, thermal camera needed to be calibrated) – Needs precise registration. • Software only : Based on Head size (Fitting) • Eye Detection – Visual • Recognition – Fusion of Visual and Thermal • Only thermal image based face recognition • Using thermal signatures based on temperature. • Eye Detection : ? ( Face Detection will be easier) • Glasses on : Needs other features • Subject cooperation needed ( When enter facilities, subjects need to be taken off glasses -> Intrusive)

More Related