Abstract: An image sensor includes: a light detector including a plurality of photosensitive cells configured to sense light; a color separation lens array provided above the light detector and including a plurality of pattern structures, the color separation lens array being configured to collect light having different wavelength spectra respectively on at least two photosensitive cells of the plurality of photosensitive cells; and a variable interlayer element configured to adjust an optical distance between the light detector and the color separation lens array.

Abstract: Provided are a method and apparatus, which, during video encoding and decoding processes, obtain chroma intra prediction mode information about a current chroma block, when the chroma intra prediction mode information indicates a direct mode (DM), determine a luma block including a luma sample corresponding to a chroma sample at a lower-right location with respect to a center of the current chroma block, determine a chroma intra prediction mode of the current chroma block based on an intra prediction mode of the determined luma block, and perform intra prediction on the current chroma block, based on the determined chroma intra prediction mode.

Abstract: A method of decoding an image according to an embodiment includes: when a size of a current block in the image is equal to or greater than a certain size, determining a candidate list including, as a candidate, a first reference block indicated by a temporal motion vector; when the first reference block is selected from among candidates included in the candidate list, determining motion vectors of sub-blocks in the current block by using motion vectors obtained from the first reference block; and reconstructing the current block based on sample values of a second reference block indicated by the motion vectors of the sub-blocks.

Abstract: A switching regulator generates an output voltage from an input voltage and includes; a charge sharing circuit that selectively forms one of a first charge sharing path between a first flying capacitor and a second bootstrap capacitor and a second charge sharing path between a second flying capacitor and a first bootstrap capacitor based on first and second conversion modes.

Abstract: In various examples, to support training a deep neural network (DNN) to predict a dense representation of a 3D surface structure of interest, a training dataset is generated using a parametric mathematical modeling. A variety of synthetic 3D road surfaces may be generated by modeling a 3D road surface using varied parameters to simulate changes in road direction and lateral surface slope. In an example embodiment, a synthetic 3D road surface may be created by modeling a longitudinal 3D curve and expanding the longitudinal 3D curve to a 3D surface, and the resulting synthetic 3D surface may be sampled to form a synthetic ground truth projection image (e.g., a 2D height map). To generate corresponding input training data, a known pattern that represents which pixels may remain unobserved during 3D structure estimation may be generated and applied to a ground truth projection image to simulate a corresponding sparse projection image.

Abstract: In various examples, live perception from sensors of a vehicle may be leveraged to generate potential paths for the vehicle to navigate an intersection in real-time or near real-time. For example, a deep neural network (DNN) may be trained to compute various outputs—such as heat maps corresponding to key points associated with the intersection, vector fields corresponding to directionality, heading, and offsets with respect to lanes, intensity maps corresponding to widths of lanes, and/or classifications corresponding to line segments of the intersection. The outputs may be decoded and/or otherwise post-processed to reconstruct an intersection—or key points corresponding thereto—and to determine proposed or potential paths for navigating the vehicle through the intersection.

Abstract: In various examples, a deep neural network (DNN) is trained to accurately predict, in deployment, distances to objects and obstacles using image data alone. The DNN may be trained with ground truth data that is generated and encoded using sensor data from any number of depth predicting sensors, such as, without limitation, RADAR sensors, LIDAR sensors, and/or SONAR sensors. Camera adaptation algorithms may be used in various embodiments to adapt the DNN for use with image data generated by cameras with varying parameters—such as varying fields of view. In some examples, a post-processing safety bounds operation may be executed on the predictions of the DNN to ensure that the predictions fall within a safety-permissible range.

Abstract: A video decoding method includes predicting a current block according to an intra prediction mode of the current block, determining whether to apply position dependent intra prediction filtering to the current block according to the intra prediction mode of the current block, when the position dependent intra prediction filtering is applied to the current block, determining at least one of an upper reference sample, a left reference sample, an upper weight, and a left weight for the position dependent intra prediction filtering of a current sample of the current block according to the intra prediction mode of the current block, and applying the position dependent intra prediction filtering to the current sample of the current block according to at least one of the upper reference sample, the left reference sample, the upper weight, and the left weight.

Abstract: Provided is a video decoding method including obtaining identification information of a first tile and identification information of a last tile from a bitstream, wherein the first and last tiles are included in a first slice, determining an index difference between the first tile and the last tile, based on a result of comparing the identification information of the first tile with the identification information of the last tile, determining the number of tiles included in the first slice, by using the index difference between the first tile and the last tile, and decoding a plurality of tiles included in the first slice according to an encoding order by using the number of tiles included in the first slice, the identification information of the first tile, and the identification information of the last tile.

Abstract: A video decoding method includes predicting a current block according to an intra prediction mode of the current block, determining whether to apply position dependent intra prediction filtering to the current block according to the intra prediction mode of the current block, when the position dependent intra prediction filtering is applied to the current block, determining at least one of an upper reference sample, a left reference sample, an upper weight, and a left weight for the position dependent intra prediction filtering of a current sample of the current block according to the intra prediction mode of the current block, and applying the position dependent intra prediction filtering to the current sample of the current block according to at least one of the upper reference sample, the left reference sample, the upper weight, and the left weight.

Abstract: A video decoding method includes predicting a current block according to an intra prediction mode of the current block, determining whether to apply position dependent intra prediction filtering to the current block according to the intra prediction mode of the current block, when the position dependent intra prediction filtering is applied to the current block, determining at least one of an upper reference sample, a left reference sample, an upper weight, and a left weight for the position dependent intra prediction filtering of a current sample of the current block according to the intra prediction mode of the current block, and applying the position dependent intra prediction filtering to the current sample of the current block according to at least one of the upper reference sample, the left reference sample, the upper weight, and the left weight.

Abstract: A video decoding method includes predicting a current block according to an intra prediction mode of the current block, determining whether to apply position dependent intra prediction filtering to the current block according to the intra prediction mode of the current block, when the position dependent intra prediction filtering is applied to the current block, determining at least one of an upper reference sample, a left reference sample, an upper weight, and a left weight for the position dependent intra prediction filtering of a current sample of the current block according to the intra prediction mode of the current block, and applying the position dependent intra prediction filtering to the current sample of the current block according to at least one of the upper reference sample, the left reference sample, the upper weight, and the left weight.

Abstract: In various examples, live perception from sensors of an ego-machine may be leveraged to detect objects and assign the objects to bounded regions (e.g., lanes or a roadway) in an environment of the ego-machine in real-time or near real-time. For example, a deep neural network (DNN) may be trained to compute outputs—such as output segmentation masks—that may correspond to a combination of object classification and lane identifiers. The output masks may be post-processed to determine object to lane assignments that assign detected objects to lanes in order to aid an autonomous or semi-autonomous machine in a surrounding environment.

Abstract: A method, performed by a decoding apparatus, of decoding motion information, including: obtaining change information indicating whether to change motion information of a neighboring block of a current block; based on the change information indicating to change the motion information, determining a prediction direction of the current block by changing a prediction direction of the neighboring block; based on the prediction direction of the neighboring block being a first uni-direction and the prediction direction of the current block being a second uni-direction or a bi-direction, selecting a reference picture of the second uni-direction of the current block from a second reference picture list; obtaining a motion vector of the second uni-direction of the current block; and reconstructing the current block using a reference block indicated by the motion vector of the second uni-direction of the current block.

Abstract: A video decoding method includes: determining a largest size of a coding unit and a smallest size of the coding unit by using information about the largest size of the coding unit and information about the smallest size of the coding unit, the information being obtained from a bitstream; obtaining, from the bitstream, information about a largest size of a block allowed to be ternary split and information about a smallest size of the block allowed to be ternary split, determining the largest size of the block allowed to be ternary split, determining the smallest size of the block allowed to be ternary split, determining whether to ternary split a current block, and decoding blocks generated by ternary splitting the current block.

Abstract: A video decoding method, including obtaining, from a bitstream, information indicating a prediction type of a coding unit including a current block; obtaining, from the bitstream, coded block flag (CBF) information for a Cr component of the current block and CBF information for a Cb component of the current block; determining a weight of a residual sample of the Cr component and a weight of a residual sample of the Cb component, based on the prediction type of the coding unit, the CBF information for the Cr component, and the CBF information for the Cb component; obtaining, from the bitstream, a chroma joint residual sample of the current block; and reconstructing the residual sample of the Cr component by using the chroma joint residual sample of the current block and the weight of the residual sample of the Cr component, and reconstructing the residual sample of the Cb component by using the chroma joint residual sample of the current block and the weight of the residual sample of the Cb component.

Abstract: Provided is an image decoding method including determining a current chroma block having a rectangular shape corresponding to a current luma block included in one of a plurality of luma blocks, determining a piece of motion information for the current chroma block and a chroma block adjacent to the current chroma block by using motion information of the current chroma block and the adjacent chroma block, and performing inter prediction on the current chroma block and the adjacent chroma block by using the piece of motion information for the current chroma block and the adjacent chroma block to generate prediction blocks of the current chroma block and the adjacent chroma block.

Abstract: In various examples, live perception from sensors of a vehicle may be leveraged to generate potential paths for the vehicle to navigate an intersection in real-time or near real-time. For example, a deep neural network (DNN) may be trained to compute various outputs—such as heat maps corresponding to key points associated with the intersection, vector fields corresponding to directionality, heading, and offsets with respect to lanes, intensity maps corresponding to widths of lanes, and/or classifications corresponding to line segments of the intersection. The outputs may be decoded and/or otherwise post-processed to reconstruct an intersection—or key points corresponding thereto—and to determine proposed or potential paths for navigating the vehicle through the intersection.

Abstract: In various examples, to support training a deep neural network (DNN) to predict a dense representation of a 3D surface structure of interest, a training dataset is generated from real-world data. For example, one or more vehicles may collect image data and LiDAR data while navigating through a real-world environment. To generate input training data, 3D surface structure estimation may be performed on captured image data to generate a sparse representation of a 3D surface structure of interest (e.g., a 3D road surface). To generate corresponding ground truth training data, captured LiDAR data may be smoothed, subject to outlier removal, subject to triangulation to filling missing values, accumulated from multiple LiDAR sensors, aligned with corresponding frames of image data, and/or annotated to identify 3D points on the 3D surface of interest, and the identified 3D points may be projected to generate a dense representation of the 3D surface structure.

Abstract: Provided is an image decoding method including determining a plurality of coding units in a chroma image by hierarchically splitting the chroma image, based on a split shape mode of blocks in the chroma image of a current image, and decoding the current image, based on the plurality of coding units in the chroma image. In this regard, the determining of the plurality of coding units in the chroma image may include, when a size or an area of a chroma block from among a plurality of chroma blocks to be generated by splitting a current chroma block in the chroma image is equal to or smaller than a preset size or a preset area, not allowing splitting of the current chroma block based on a split shape mode of the current chroma block, and determining at least one coding unit included in the current chroma block.