Abstract: Provided is an image decoding method including: obtaining, from a bitstream, a syntax element regarding multiple transform selection (MTS) with respect to a current coding unit or a current transform unit included in the current coding unit; determining a horizontal transform kernel or a vertical transform kernel with respect to the current transform unit based on the obtained syntax element; obtaining a residual signal by performing inverse transformation on the current transform unit, based on the determined horizontal transform kernel or vertical transform kernel with respect to the current transform unit; and generating a reconstruction block including the current coding unit or the current transform unit based on the residual signal with respect to the current transform unit.

Abstract: Provided is an image decoding method including: obtaining, from a bitstream, a syntax element regarding multiple transform selection (MTS) with respect to a current coding unit or a current transform unit included in the current coding unit; determining a horizontal transform kernel or a vertical transform kernel with respect to the current transform unit based on the obtained syntax element; obtaining a residual signal by performing inverse transformation on the current transform unit, based on the determined horizontal transform kernel or vertical transform kernel with respect to the current transform unit; and generating a reconstruction block including the current coding unit or the current transform unit based on the residual signal with respect to the current transform unit.

Abstract: Provided is an image decoding method including: obtaining, from a bitstream, a syntax element regarding multiple transform selection (MTS) with respect to a current coding unit or a current transform unit included in the current coding unit; determining a horizontal transform kernel or a vertical transform kernel with respect to the current transform unit based on the obtained syntax element; obtaining a residual signal by performing inverse transformation on the current transform unit, based on the determined horizontal transform kernel or vertical transform kernel with respect to the current transform unit; and generating a reconstruction block including the current coding unit or the current transform unit based on the residual signal with respect to the current transform unit.

Abstract: In various examples, systems and methods are disclosed that preserve rich spatial information from an input resolution of a machine learning model to regress on lines in an input image. The machine learning model may be trained to predict, in deployment, distances for each pixel of the input image at an input resolution to a line pixel determined to correspond to a line in the input image. The machine learning model may further be trained to predict angles and label classes of the line. An embedding algorithm may be used to train the machine learning model to predict clusters of line pixels that each correspond to a respective line in the input image. In deployment, the predictions of the machine learning model may be used as an aid for understanding the surrounding environment—e.g., for updating a world model—in a variety of autonomous machine applications.

Abstract: An image decoding method may include obtaining a first coded block flag, when the first coded block flag of the current coding unit indicates that the current coding unit comprises the one or more non-zero significant transform coefficients, identifying whether at least one of a height and a width of the current coding unit is greater than a predetermined size, based on whether the at least one of the height and the width of the current coding unit is greater than the predetermined size, obtaining at least one transform unit, when the at least one of the height and the width of the current coding unit is greater than the predetermined size, obtaining a second coded block flag, obtaining a residual signal of the block of the luma component based on the second coded block flag, and reconstructing the current coding unit based on the residual signal.

Abstract: In various examples, sensor data representative of an image of a field of view of a vehicle sensor may be received and the sensor data may be applied to a machine learning model. The machine learning model may compute a segmentation mask representative of portions of the image corresponding to lane markings of the driving surface of the vehicle. Analysis of the segmentation mask may be performed to determine lane marking types, and lane boundaries may be generated by performing curve fitting on the lane markings corresponding to each of the lane marking types. The data representative of the lane boundaries may then be sent to a component of the vehicle for use in navigating the vehicle through the driving surface.

Abstract: A video decoding method includes determining whether an ultimate motion vector expression (UMVE) mode is allowed for an upper data unit including a current block, when the UMVE mode is allowed for the upper data unit, determining whether the UMVE mode is applied to the current block, when the UMVE mode is applied to the current block, determining a base motion vector of the current block, determining a correction distance and a correction direction for correction of the base motion vector, determining a motion vector of the current block by correcting the base motion vector according to the correction distance and the correction direction, and reconstructing the current block based on the motion vector of the current block.

Abstract: In various examples, a three-dimensional (3D) intersection structure may be predicted using a deep neural network (DNN) based on processing two-dimensional (2D) input data. To train the DNN to accurately predict 3D intersection structures from 2D inputs, the DNN may be trained using a first loss function that compares 3D outputs of the DNN—after conversion to 2D space—to 2D ground truth data and a second loss function that analyzes the 3D predictions of the DNN in view of one or more geometric constraints—e.g., geometric knowledge of intersections may be used to penalize predictions of the DNN that do not align with known intersection and/or road structure geometries. As such, live perception of an autonomous or semi-autonomous vehicle may be used by the DNN to detect 3D locations of intersection structures from 2D inputs.

Abstract: An image processing method and an image processing apparatus are provided to obtain input data for deblocking filtering based on deblocking filtering target pixels of at least one line perpendicular to a boundary line of blocks and encoding information about the deblocking filtering target pixels of at least one line, obtain a feature map of the input data by inputting the input data to a first neural network, obtain a deblocking filter coefficient by inputting the feature map to a second neural network, obtain a deblocking filter compensation value by inputting the feature map to a third neural network, obtain a deblocking filter strength by inputting the input data to a fourth neural network, obtain deblocking filtered pixels by performing deblocking filtering on the deblocking filtering target pixels of the at least one line using the deblocking filter coefficient, the deblocking filter compensation value, and the deblocking filter strength.

Abstract: A display apparatus including pixels arranged in a display area includes a first conductive layer including a first voltage line, a second conductive layer disposed on the first conductive layer and including a first conductive pattern overlapping the first voltage line, a semiconductor layer disposed on the second conductive layer and including a first semiconductor pattern overlapping the first conductive pattern, a third conductive layer disposed on the semiconductor layer and including a second conductive pattern overlapping the first conductive pattern, and a fourth conductive layer disposed on the third conductive layer and including a data line and a third conductive pattern overlapping the second conductive pattern, wherein the first voltage line includes a body portion extending in a first direction and a shielding portion extending from the body portion in a second direction to overlap the data line, wherein the second direction crosses the first direction.

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: determining a prediction mode of a current block to be one of a skip mode and a merge mode; when a motion vector, which is determined from a merge candidate list of the current block, and a merge motion vector difference are to be used, obtaining merge candidate information indicating one candidate in the merge candidate list by performing entropy encoding on a bitstream by applying one piece of context information; determining a base motion vector from one candidate determined from the merge candidate list, based on the merge candidate information; and determining a motion vector of the current block by using a distance index of a merge motion vector difference of the current block and a direction index of the merge motion vector difference to use the base motion vector and the merge motion vector difference.

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: Provided is an image decoding method including: determining a first coding block and a second coding block corresponding to the first coding block; when a size of the first coding block is equal to or smaller than a preset size, obtaining first split shape mode information and second split shape mode information from a bitstream; determining a split mode of the first coding block, based on the first split shape mode information, and determining a split mode of the second coding block, based on the second split shape mode information; and decoding a coding block of a first color component which is determined based on the split mode of the first coding block and a coding block of a second color component which is determined based on the split mode of the second coding block.

Abstract: In various examples, sensor data representative of an image of a field of view of a vehicle sensor may be received and the sensor data may be applied to a machine learning model. The machine learning model may compute a segmentation mask representative of portions of the image corresponding to lane markings of the driving surface of the vehicle. Analysis of the segmentation mask may be performed to determine lane marking types, and lane boundaries may be generated by performing curve fitting on the lane markings corresponding to each of the lane marking types. The data representative of the lane boundaries may then be sent to a component of the vehicle for use in navigating the vehicle through the driving surface.

Abstract: A display panel which includes a base layer, a pixel defining layer on the base layer and that has a light emitting opening, a barrier wall on the pixel defining layer and that has a barrier wall opening that overlaps the light emitting opening, and a light emitting element including an anode, an intermediate layer, and a cathode that makes contact with the barrier wall. The barrier wall opening includes a first inner surface and a third inner surface that extend in a direction parallel to a first direction, a second inner surface and a fourth inner surface that extend in a direction parallel to a second direction crossing the first direction, a first corner region, a second corner region, a third corner region, and a fourth corner region.

Abstract: An image decoding method may include: receiving a bitstream generated as a result of encoding an image sequence; obtaining, from a sequence parameter set of the bitstream, a first tool set index indicating a tool allowed to decode the bitstream among a plurality of tools; obtaining, from the sequence parameter set, tool flags based on the tool flags, identifying a tool that has been used to encode the image sequence among the plurality of tools; and reconstructing the image sequence based on the identified tool, wherein values of the tool flags are set according to a value of the first tool set index.

Abstract: In various examples, a 3D surface structure such as the 3D surface structure of a road (3D road surface) may be observed and estimated to generate a 3D point cloud or other representation of the 3D surface structure. Since the estimated representation may be sparse, a deep neural network (DNN) may be used to predict values for a dense representation of the 3D surface structure from the sparse representation. For example, a sparse 3D point cloud may be projected to form a sparse projection image (e.g., a sparse 2D height map), which may be fed into the DNN to predict a dense projection image (e.g., a dense 2D height map). The predicted dense representation of the 3D surface structure may be provided to an autonomous vehicle drive stack to enable safe and comfortable planning and control of the autonomous vehicle.

Abstract: A memory device includes a memory cell including a selection layer and a phase change material layer, and a controller, wherein the selection layer includes a switching material, the phase change material layer includes a phase change material, and the controller is configured to apply a write pulse to the selection layer and the phase change material layer and control a polarity, a peak value, and a shape of the write pulse.

Abstract: In various examples, systems and methods are disclosed that preserve rich spatial information from an input resolution of a machine learning model to regress on lines in an input image. The machine learning model may be trained to predict, in deployment, distances for each pixel of the input image at an input resolution to a line pixel determined to correspond to a line in the input image. The machine learning model may further be trained to predict angles and label classes of the line. An embedding algorithm may be used to train the machine learning model to predict clusters of line pixels that each correspond to a respective line in the input image. In deployment, the predictions of the machine learning model may be used as an aid for understanding the surrounding environment—e.g., for updating a world model—in a variety of autonomous machine applications.