1 / 13

Spectral LWIR Imaging for Remote Face Detection

Spectral LWIR Imaging for Remote Face Detection. Dalton Rosario U.S. Army Research Laboratory IEEE IGARSS, Vancouver, Canada 29 July 2011. Outline. Unrelated Operational Concept A Difficult Target Detection Problem Proposed Algorithmic Framework Experimental Results

jorryn
Télécharger la présentation

Spectral LWIR Imaging for Remote Face Detection

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. Spectral LWIR Imaging for Remote Face Detection Dalton Rosario U.S. Army Research Laboratory IEEE IGARSS, Vancouver, Canada 29 July 2011

  2. Outline • Unrelated Operational Concept • A Difficult Target Detection Problem • Proposed Algorithmic Framework • Experimental Results • Adaptation to LWIR Specific-Face Detection • Experimental Results • Concluding Remarks

  3. Operational Scenarios Target Visible-NIR-SWIR 320 x 256 x 225

  4. Some Comments • Non-kinematic based target detection/ tracking • Advantages Using Hyperspectral Imagery • No geo-rectification required • No frame-to-frame registration required • Target detection (moving or stationary) • Handles challenges in kinematic based methods • Challenge • Subset of Curse of Dimensionality Problem • Atmospheric variation, geometry of illumination, etc • Kinematic based methods • Challenges • Changes in velocity • Proximity to other vehicles • Prolonged obscuration

  5. A Fundamental Problem & A Solution Problem Contrast Contrast

  6. Algorithmic Concept Framework

  7. Proof of Principle ExperimentSpectral Tracking –Frame i Pseudo-Color Target

  8. Proof of Principle ExperimentSpectral Tracking –Frame i+1

  9. Proof of Principle ExperimentSpectral Tracking –Frame i+40

  10. Target

  11. LWIRHyperspectral Specific Face Detection Contrast Contrast • Assumptions: • Range is known • Facial spectral mixture is distinct LWIR 8-11 mm 410 bands 400 ft 300 ft 200 ft Unknown Probability Distribution Functions

  12. LWIRHyperspectral Specific Face Detection Pseudo-Color 200 ft 300 ft 400 ft Target • Algorithm Suite • First Level of Detection • Temperature & Emissivity Separation. • Use human body biometrics for Skin detection • Uniform Temperature (35.5 to 37.5 oC) • IR Emissivity relatively uniform among different skin • Second Level – Specific Face Detection • Apply All bands Statistical Hypothesis Test Afterward

  13. Concluding Remarks Introduced an algorithmic framework for extremely small sample size multivariate target detection problems (n << B) Approach is Flexible, Adaptive Approach Addresses Fusion of Spectral Regions Visible, NIR, SWIR, MWIR, LWIR Proof of principle experimentation for LWIR Specific-Human-Face Detection First Level Detection: Human skin biometrics (temperature & emissivity ranges) Second Level – Proposed approach using All Bands on candidate regions from first level

More Related