Konstantin M. Yemelyanov

Polarization-based and specular-reflection-based noncontact latent fingerprint imaging and lifting

In forensic science the finger marks left unintentionally by people at a crime scene are referred to as latent fingerprints. Most existing techniques to detect and lift latent fingerprints require application of a certain material directly onto the exhibit. The chemical and physical processing applied to the fingerprint potentially degrades or prevents further forensic testing on the same evidence sample. Many existing methods also have deleterious side effects. We introduce a method to detect and extract latent fingerprint images without applying any powder or chemicals on the object. Our method is based on the optical phenomena of polarization and specular reflection together with the physiology of fingerprint formation. The recovered image quality is comparable to existing methods. In some cases, such as the sticky side of tape, our method shows unique advantages.

Separation and contrast enhancement of overlapping cast shadow components using polarization

Shadow is an inseparable aspect of all natural scenes. When there are multiple light sources or multiple reflections several different shadows may overlap at the same location and create complicated patterns. Shadows are a potentially good source of information about a scene if the shadow regions can be properly identified and segmented. However, shadow region identification and segmentation is a difficult task and improperly identified shadows often interfere with machine vision tasks like object recognition and tracking. We propose here a new shadow separation and contrast enhancement method based on the polarization of light. Polarization information of the scene captured by our polarization-sensitive camera is shown to separate shadows from different light sources effectively. Such shadow separation is almost impossible to realize with conventional, polarization-insensitive imaging.

Adaptive algorithms for two-channel polarization sensing under various polarization statistics with nonuniform distributions

The polarization of light carries much useful information about the environment. Biological studies have shown that some animal species use polarization information for navigation and other purposes. It has been previously shown that a bioinspired polarization-difference imaging (PDI) technique can facilitate detection and feature extraction of targets in scattering media. It has also been established [J. Opt. Soc. Am. A 15, 359 (1998)] that polarization sum and polarization difference are the optimum pair of linear combinations of images taken through two orthogonally oriented linear polarizers of a scene having a uniform distribution of polarization directions. However, in many real environments the scene has a nonuniform distribution of polarization directions. Using principal component analysis of the polarization statistics of the scene, we develop a method to determine the two optimum information channels with unequal weighting coefficients that can be formed as linear combinations of the images of a scene taken through a pair of linear polarizers not constrained to the horizontal and vertical directions of the scene. We determine the optimal orientations of linear polarization filters that enhance separation of a target from the background, where the target is defined as an area with distinct polarization characteristics as compared to the background. Experimental results confirm that in most situations adaptive PDI outperforms conventional PDI with fixed channels.

Bio-inspired display of polarization information using selected visual cues

For imaging systems the polarization of electromagnetic waves carries much potentially useful information about such features of the world as the surface shape, material contents, local curvature of objects, as well as about the relative locations of the source, object and imaging system. The imaging system of the human eye however, is “polarization-blind”, and cannot utilize the polarization of light without the aid of an artificial, polarization-sensitive instrument. Therefore, polarization information captured by a man-made polarimetric imaging system must be displayed to a human observer in the form of visual cues that are naturally processed by the human visual system, while essentially preserving the other important non-polarization information (such as spectral and intensity information) in an image. In other words, some forms of sensory substitution are needed for representing polarization “signals” without affecting other visual information such as color and brightness. We are investigating several bio-inspired representational methodologies for mapping polarization information into visual cues readily perceived by the human visual system, and determining which mappings are most suitable for specific applications such as object detection, navigation, sensing, scene classifications, and surface deformation. The visual cues and strategies we are exploring are the use of coherently moving dots superimposed on image to represent various range of polarization signals, overlaying textures with spatial and/or temporal signatures to segregate regions of image with differing polarization, modulating luminance and/or color contrast of scenes in terms of certain aspects of polarization values, and fusing polarization images into intensity-only images. In this talk, we will present samples of our findings in this area.

Polarization enhanced visual surveillance techniques

Visual surveillance relies on visible light to convey information about the environment. An important but less utilized characteristic of light in surveillance is its polarization due to the transverse wave nature of electromagnetic waves. Although the predominant natural light source, the sun light, appear to be unpolarized, the light that actually enters a surveillance camera often has biased polarization distribution due to scattering and reflection by the environment. These differently biased polarization patterns thus carry important information about the environment, such as material type, object shape ...etc. With proper optical setup and computer vision analysis, polarization information can be captured and used to enhance traditional polarization-insensitive visual surveillance techniques. This work presents polarization based optical and computational techniques to improve target detection, to perform novel optical navigation, and to help resolve ambiguities in traditional visual sensors.

Adaptive polarization-difference imaging algorithms for through-the-wall microwave imaging scenarios

The preliminary results of application of the adaptive polarization-imaging algorithm for through-the-wall microwave imaging problems are presented. Use of complete polarization information in the scattered field from the object together with the adaptation technique provides enhancement in detection of target movement.

Polarization-Based Segmentation for Enhancement of Target Detection in Adaptive Polarization-Difference Imaging

The polarization-based segmentation algorithm, which is an extension of our adaptive polarization-difference technique, has been developed. Notable increase in target-against-background separation is achieved.

Optical non-invasive latent fingerprint lifting

Most existing latent fingerprints lifting methods are invasive and often induce deleterious side effects. Based on the optical reflection and polarization we present a completely non-invasive method to lift latent fingerprints into good quality images.

Shadow vision enhancement using polarization imaging

Shadow is omnipresent in natural scenes. Conventional imaging methodologies are less effective in detecting features in shadow regions. Polarization imaging exploits another dimension of light that increases detection sensitivity in shadow and reveals hidden features.

Bio-inspired visualization of polarization information using temporal fusion, flicker, coherently moving dots and texture

The human eye is effectively “polarization-blind”, i.e., it is only sensitive to the intensity and wavelength of optical waves, and encodes and maps these into the perceptual qualities of brightness and color. Polarization, the third characteristic of optical waves, however, has been shown to be an important information-carrying features of waves, and may reveal useful information about a scene such as local curvature, material contrast, surface shapes, and the relative direction of illumination [1]. Since the unaided human eye cannot sense the polarization of light, information collected by polarimetric imaging systems must be exh ibited as a set of cues visible to a human observer, preferably in a way that, by and large, preserves important non-polarization information such as that carried by the spectral and luminance distributions in the image. Effectively, some form of “sensory substitution” is required for presenting polarization “signals” to a “polarization-blind” observer, without affecting other visual information such as color and brightness.