Lumidigm, Inc.

Hid Global Corporation

Assa Abloy Ab

Ryan Eric Martin

Robert Kjell Rowe

Stephen P Corcoran

Horst Arnold Mueller

Kristin A Nixon

Todd Doucet

Steve Corcoran

Gary Rogers

Nathaniel I Matter

Alex Litz

Disclosed herein are systems and methods for using machine learning for image-based spoof detection. One embodiment takes the form of a method that includes obtaining an input-data set that includes a plurality of images captured of a biometric-authentication subject by a plurality of cameras of a camera system. The method also includes inputting the input-data set into a trained machine-learning module, and processing the input-data set using the machine-learning module to obtain, from the machine-learning module, a spoof-detection result for the biometric-authentication subject. The method also includes outputting the spoof-detection result for the biometric-authentication subject.

Systems and methods for using machine learning for image-based spoof detection

Systems and methods are described for providing reliable biometric access control using an optical biometric sensor in a miniaturized form factor. Some implementations include multiple light sources that can illuminate skin or other tissue at multiple locations during a single measurement session. An imaging array can be arranged to form images of the light exiting the tissue only after undergoing diffuse reflectance in the tissue. Some implementations use the images to perform biometric functions. For example, the images can be used to identify an individual, verify identity of an individual, estimate demographic characteristics of an individual, etc. Such biometric functions can further be used to determine and affect access to secured assets.

Miniaturized optical biometric sensing

A variety of forms of machine-readable symbols are disclosed, as well as methods and systems of constructing machine-readable symbols, methods and systems of acquiring machine-readable symbols, and methods and systems of decoding machine-readable symbols.

Machine-readable symbols

Methods and devices of studying a predefined portion of an object having a feature of interest are disclosed. The feature of interest defines a class of objects that includes the object. Light sources directly illuminate the object from different illumination directions. The light sources are maintained in a stable configuration relative to the object. For each illumination direction, an image is generated from light scattered from the object with a camera maintained in a stable configuration relative to the light sources. A methodology derived from machine learning for the class of objects is applied to filter the generated images are filtered for a characteristic consistent with the feature of interest. Surface gradients are determined from the filtered images and integrated to generate a topography of a surface of the object.

Optical topographic imaging

Methods and sensors are disclosed for executing a biometric function. Illumination light is generated and directed to a deformable layer when the deformable layer is in a state of deformation resulting from pressure applied to the deformable layer by a skin site. Light scattered from the deformable layer is received. A fingerprint pattern of the skin site is determined from the received light. The biometric function is performed with the determined fingerprint pattern.

Methods and systems for performing biometric functions

Embodiments disclose a two imager biometric sensor. In some embodiments, the two imagers can include a direct imager and a TIR imager. In some embodiments, multispectral light sources can be used to illuminate target tissue imaged by two imagers. In some embodiments, composite images can be created from images detected using both imagers.

Dual-imager biometric sensor

Some embodiments of the present invention disclose systems and methods for a multiplexed multispectral imaging, object discrimination, background discrimination, and/or object identification. In some embodiments, a multispectral sensor is provided that includes at least two illumination sources and an imager with a color filter array. The two illumination sources can illuminate a platen with light having distinct illumination angles as well as distinct characteristics, such as distinct wavelengths or wavelength bands and/or distinct polarization conditions. The color filter array, which may be integral with the imager or separate therefrom, can filter light based on the specific distinctions between the two illumination sources. A single image of an object at the platen can be acquired. Individual pixels of the image will then be highly associated with one or the other illumination source. Because of the filtering by the color filter array, this image will include information about the illumination angle.

Multiplexed biometric imaging

A biometric feature detection system is provided according to one embodiment. The system may include a target, an illumination source, a color filter array, a light detector, and a computational unit. The illumination source illuminates a portion of the target with monochromatic light at a large angle of incidence measured from the normal of the target surface. The light detector may be configured to receive light from the target through the color-filter array and provide an image of the target surface. Each pixel of the light detector may correspond to one of a plurality of color filter mosaics such that each pixel detects light associated with a corresponding color filter. The computational unit may be interfaced with at least the light detector. The computational unit may include instructions to monitor the levels of monochromatic light and make proximity, presence and segmentation determinations based on the detected levels of monochromatic light.

Biometric detection using spatial, temporal, and/or spectral techniques

Methods and systems are provided for performing a biometric function. A purported skin site of an individual is illuminated with white light. Light scattered from the purported skin site is received with a color imager on which the received light is incident. Spatially distributed images of the purported skin site are derived and correspond to different volumes of illuminated tissue of the individual. The images are analyzed to perform the biometric function.

White-light spectral biometric sensors

Combined multispectral and total-internal-reflectance biometric imaging systems are disclosed. A platen has multiple facets, at least one of which has a surface adapted for placement of a purported skin site by an individual and another facet may include an optical absorber. An illumination source and an optical arrangement are disposed to illuminate the purported skin site with light from the illumination source along distinct illumination paths, including paths at angles less than the critical angle and paths at angles greater than the critical angle. Both multispectral and total-internal-reflectance illumination are received by an imaging system. The imaging system may include first and second imaging locations adapted to record images from separate illumination paths. The platen may also include non parallel exits facets.

Growing community of inventors

Tijeras, NM, United States of America

Ryan Eric Martin