No abstract

We present a series of long-wave-infrared (LWIR) polarimetric-based thermal images of facial profiles in which polarization-state information of the image-forming radiance is retained and displayed. The resultant polarimetric images show enhanced facial features, additional texture, and details that are not present in corresponding conventional thermal imagery. It has been generally thought that conventional thermal imagery (MidIR or LWIR) could not produce the detailed spatial information required for reliable human identification due to the so-called "ghosting" effect often seen in thermal imagery of human subjects. By using polarimetric information, we are able to extract subtle surface features of the human face, thus improving subject identification. Polarimetric image sets considered include the conventional thermal intensity image, S0, the two Stokes images, S1 and S2, and a Stokes image product called the degree-of-linear-polarization image.

The mass density normalized absorption and total scattering coefficients have been measured using in situ techniques at selected wavelengths from the visible to approximately 1 cm for soot generated by the open combustion of diesel fuel. Particle morphologies are complex although similar to those of soots of other hydrocarbons and methods of generation. An ellipsoidal model has been applied as an approximation to the often multiconnected, chainlike aerosol and then compared with the measured results. The experimental results show an approximate (lambda)(-1) dependence over more than five decades of wavelength data. There is only general agreement with the simplified calculations in this feature as well as in the magnitude.

We investigate the performance of polarimetric imaging in the long-wave infrared (LWIR) spectrum for cross-modal face recognition. For this work, polarimetric imagery is generated as stacks of three components: the conventional thermal intensity image (referred to as S₀), and the two Stokes images, S₁ and S₂, which contain combinations of different polarizations. The proposed face recognition algorithm extracts and combines local gradient magnitude and orientation information from S₀, S₁, and S₂ to generate a robust feature set that is well-suited for cross-modal face recognition. Initial results show that polarimetric LWIR-to-visible face recognition achieves an 18% increase in Rank-1 identification rate compared to conventional LWIR-to-visible face recognition. We conclude that a substantial improvement in automatic face recognition performance can be achieved by exploiting the polarization-state of radiance, as compared to using conventional thermal imagery.