Abstract: A light field imaging device may include a diffraction grating assembly configured to receive an optical wavefront from a scene and including one or more diffraction gratings. Each diffraction grating has a refractive index modulation pattern with a grating period along a grating axis and is configured to generate a diffracted wavefront. The device may also include a pixel array configured to detect the diffracted wavefront in a near-field region. The pixel array includes light-sensitive pixels and a pixel pitch along the grating axis that is equal to or larger than the grating period. Each pixel samples a portion of the diffracted wavefront and generates a pixel response. The pixels include groups or pairs of adjacent pixels, where the adjacent pixels in each group or pair have different pixel responses as a function of the angle of incidence of the optical wavefront. Light field imaging methods are also disclosed.

Abstract: A method for spoofing detection in a biometric object can include receiving an image pair representing two different viewpoints of the biometric object, and inputting the image pair into a disparity encoder of a trained neural network configured to map the image pair to an encoded signal conveying disparity information associated with the image pair, wherein the disparity encoder was trained for disparity map estimation together with a disparity decoder, the disparity encoder and the disparity decoder forming a disparity branch of the neural network during training. The method can also include inputting the encoded signal into a classification branch of the trained neural network configured to map the encoded signal to a probability that the biometric object is genuine or spoofed, wherein the classification branch was trained for spoofing detection together with the disparity branch. Techniques of training a neural network to perform image-based biometric anti-spoofing are also disclosed.

Abstract: A method of managing power usage of a depth imaging system including an image sensor having an array of pixels configured to detect light incident from a scene and an optical encoder configured to modulate the incident light detected by the pixels in accordance with an angle of incidence of the incident light. The method includes operating the system in a lower power mode corresponding to a first power consumption level, including capturing, with the pixels, image data of the scene having angle-dependent information encoded therein by the optical encoder, and identifying, based on the angle-dependent information, signature information in the image data indicative of a detection of an object within a specified depth range. The method also includes, in response to identifying the signature information, transitioning to operating the system in a higher power mode corresponding to a second power consumption level higher than the first power consumption level.

Abstract: An imaging system includes a transmissive diffraction mask (TDM), a pixel array, and a processor. The TDM includes a first and a second diffraction grating configured to diffract light received from a scene to generate first diffracted light encoding information about an angle of incidence (AOI) of the received light and second diffracted light encoding information about the AOI and a state of polarization (SOP) of the received light, respectively. The pixel array includes a first and a second set of pixels configured to detect the first and second diffracted light and generate therefrom a corresponding first and second set of pixel responses, respectively. The processor is configured to determine, from the first set of pixel responses, AOI data conveying the AOI of the received light, and determine, from the second set of pixel responses and the AOI data, polarization data conveying the SOP of the received light.

Abstract: Techniques for capturing three-dimensional image data of a scene and processing light field image data obtained by an optical wavefront sensor in 3D imaging applications are provided. The disclosed techniques provide a depth map of an observable scene from light field information about an optical wavefront emanating from the scene, and make use of color filters forming a color mosaic defining a primary color and one or more secondary colors, and color radial transfer functions calibrated to provide object distance information from the spatio-spectrally sampled pixel data.