Abstract: A high dynamic range sensor, comprising a pixel array, having a plurality of pixels, a spatial light transmittance modulator, the spatial light transmittance modulator having a plurality of pre-defined shadings, the plurality of pre-defined shadings are fixed with respect to the plurality of pixels, wherein the plurality of pre-defined shadings is located above the plurality of pixels and a micro-lens array, having a plurality of micro-lens aligned to at least one of the plurality of pixels, wherein the micro-lens array is located above the plurality of pixels.

Abstract: An example operation includes one or more of, determining an initial position and at least one vehicle constraint, predetermining a distance to a new position, determining a minimum turn radius based on the at least one vehicle constraint, selecting a vehicle turn radius that is at least as large as the minimum turn radius, determining a candidate position based on the at least one vehicle constraint and the vehicle turn radius, verifying whether the candidate position is valid with respect to the new position and converting the new position to the candidate position.

Abstract: A method to reduce a motion blur and a rolling shutter effect, comprising, receiving a main image frame, a main timing and a main exposure from a main camera, receiving a secondary image frame, a secondary timing and a secondary exposure from a secondary camera, correcting the secondary image frame to the main image frame, determining a delta timing based on the main timing and the secondary timing, determining a delta exposure based on the main exposure and the secondary exposure, determining a discrete motion of offset sequences based on the secondary image frame, determining a row-wise motion blur kernel based on the main image frame and the discrete motion of offset sequences and determining a spatially varying kernel deconvolution based on the main image frame and the row-wise motion blur kernel.

Abstract: A method of subgraph tile fusion in a convolutional neural network, including partitioning a network into at least one subgraph node, determining a layer order of at least one layer of the at least one subgraph node, determining a input layer of the at least one subgraph node, determining a weight layer of the at least one subgraph node, determining a output layer of the at least one subgraph node and fusing the at least one subgraph node, the input layer, the weight layer and the output layer in the layer order.

Abstract: A method of camera localization, comprising, receiving a camera image, receiving a LiDAR point cloud, estimating an initial camera pose for the camera image, sampling an initial set of depth projections within the LiDAR point cloud, measuring a similarity of the initial camera pose to the initial set of depth projections and deriving a subsequent set of depth projections based on the measured similarity.

Abstract: A method of hierarchical structuring a multi-level memory in a convolutional neural network, includes partitioning a memory into a plurality of sections, partitioning the plurality of sections into a plurality of stripes, utilizing input data from the plurality of stripes in a MAC array, outputting an intermediate result from the MAC array to at least one of the plurality of stripes of a result buffer, looping back the intermediate result from the at least one of the plurality of stripes of the result buffer to at least one of the plurality of stripes of an input data buffer and outputting a final result from the at least one of the plurality of stripes of the result buffer to at least one of the plurality of stripes of an output buffer.

Abstract: A network device accesses a server via a network to collect information from a user and to verify identification of the user. Financial information of the user is accessed from one or more sources other than the user. An indicator of financial responsibility is assigned to the user based on the collected information and the accessed financial information of the user. The indicator of financial responsibility is disclosed to other persons under conditions selected by the user by communicating the indicator of financial responsibility via the network to one or more network devices.

Abstract: An example method of image signal processing, comprising at least one of, receiving a set of high priority signals, receiving a set of low priority signals, reconfiguring a first portion of a pipeline to route the high priority signals through an in-line mode process and reconfiguring a second portion of the pipeline to route the low priority signals through an offline mode process.

Abstract: The invention discloses a color balance method and device, an on-board equipment and a storage medium. The color balance method comprises the following steps of: acquiring a plurality of images of a plurality of regions in different directions around a vehicle from a plurality of cameras installed on the vehicle, wherein adjacent regions in the plurality of regions overlap each other; determining overlapping areas of the plurality of images based on overlapping areas of the adjacent regions; calculating an adjustment coefficient of the plurality of images on a color channel, and judging whether the adjustment coefficient is valid or not according to a preset threshold value; if the adjustment coefficient is invalid, setting the adjustment coefficient as a default value; and then adjusting the plurality of images according to the adjustment coefficient.

Abstract: A method of multiple-exposure multiple-frame image capture, comprising, capturing a first group of frames at an underexposed setting, selecting a first key frame from the first group of frames, aligning the first group of frames to the first key frame, merging the first group of frames into one first frame, capturing a second group of frames at the underexposed setting, selecting a second key frame from the second group of frames, aligning the second group of frames to the second key frame, merging the second group of frames into one second frame, selecting a main group of frames based on the first key frame and the second key frame, aligning the first key frame and the second key frame based on the main group of frames and merging the first key frame and the second key frame based on the main group of frames.

Abstract: A method of focusing dual cameras including receiving a fixed focus image from a fixed focus camera module, receiving an auto focus image from an auto focus camera module, calibrating a lens distortion of the auto focus camera module, calibrating a geometric relation of the auto focus camera module and the fixed focus camera module, calculating a depth of focus difference between the fixed focus image and the auto focus image, estimating an auto focus position based on the depth of focus difference and setting the auto focus position based on the estimation.

Abstract: A combination of knowledge representation and retrieval, assessments and personalization for a tutoring platform that provides an automated phased word exposition method for vocabulary learning and enable learners to learn words (or any learning objective) with sufficient proficiency and without excessive repetition. The phased learning objective method implements a moving average model that models the learner across multiple dimensions (e.g. reading, listening, writing, speaking) such that the platform is confident of the learner's understanding, while using minimal assessment doses, and such that the understanding transfers across experiences. The method implemented provides for multiple learning and assessment activities for a learning objective and leverages the same learner model for each learning dimension. Each word of a word set passes through learning and assessment-only phases for statistically significant evidence of learning.

Abstract: A method of depth estimation utilizing heterogeneous cameras, comprising, homogenizing a first camera image and a second camera image based on a first camera calibration dataset and a second camera calibration dataset respectively, wherein the first camera image and second camera image are distortion corrected and are zoom compensated, determining an initial image pair rectification transformation matrix of the homogenized first camera image and the homogenized second camera image, determining a delta image pair rectification transformation matrix based on the initial image pair rectification transformation matrix, determining a final image pair rectification transformation matrix based on the initial image pair rectification transformation matrix and the delta image pair rectification transformation matrix resulting in a final rectified image pair and disparity mapping the final rectified image pair based on a depth net regression.

Abstract: Personalized video interjections based on a learner model and a learning objective. A method for adding interjections to a video includes analyzing the content of a plurality of videos based on a set of learning objectives, selecting a video based on a learning objective, determining types of video interjections using an analytics engine that compares a learner model and the learning objective, determining a location for the video interjections using the analytics engine, generating a video interjection for each video interjection type and inserting the video interjections into the video at the determined locations.

Abstract: A method of calibration of a dual camera system, including, providing a chart of a plurality of circular rings and at least one disk, imaging the chart with a first camera to provide a first chart image and a second camera to provide a second chart image, performing edge detection of the first and second chart image on the plurality of circular rings and the at least one disk, determining the relative position of the plurality of circular rings to the at least one disk, matching the relative positions of the plurality of circular rings of the first chart image and the second chart image and aligning the first camera and the second camera based on the matching of the first chart image to the second chart image.

Abstract: A method of high precision object location in a parking lot including capturing via an entry stereo camera a template image and an entry video of a vehicle entering a parking lot, measuring a distance versus time of the vehicle entering the parking lot based on the entry video, line scanning via at least one LIDAR an outline versus time of the vehicle entering the parking lot, constructing a scanned image based on the measured distance versus time and the outline versus time and forming a three dimensional construct of the vehicle based on the template image and the scanned image.

Abstract: A three camera alignment system, having a first camera module having a first fixed zoom, a second camera module having a second fixed zoom coupled to the first camera module, wherein the second camera module is calibrated to the first camera module and a third camera module having a third fixed zoom coupled to at least one of the first camera module and the second camera module and wherein the third camera module is aligned to the calibrated first camera module and second camera module based on block matching along epipolar lines.

Abstract: A method of calibration of a dual camera system, including, providing a chart of a plurality of circular rings and at least one disk, imaging the chart with a first camera to provide a first chart image and a second camera to provide a second chart image, performing edge detection of the first and second chart image on the plurality of circular rings and the at least one disk, determining the relative position of the plurality of circular rings to the at least one disk, matching the relative positions of the plurality of circular rings of the first chart image and the second chart image and aligning the first camera and the second camera based on the matching of the first chart image to the second chart image.

Abstract: A method of distant on-road object detection, comprising, capturing an initial frame of a roadway lane in a direction of travel, detecting lane edges in the initial frame, capturing a subsequent frame of the roadway lane, cropping the subsequent frame based on the detected lane edges of the initial frame resulting in an image patch, detecting an image patch object within the image patch, down-sampling the subsequent frame, detecting a down-sampled object within the down-sampled subsequent frame and merging the detections within the image patch and the detections within the down-sampled subsequent frame.

Abstract: A method of instance segmentation, comprising, defining a bounding box, depth mapping the bounding box, sorting the depth map, estimating stability due to noise of the sorted depth map, detecting low frequency jumps of the sorted depth map based on the estimated stability, detecting high frequency jumps of the sorted depth map, comparing the detected low frequency jumps and the detected high frequency jumps, detecting at least one of a largest object and a closest object based on the comparison and pixel masking the at least one of the largest object and the closest object.