Abstract: Methods, systems, and devices for coefficient scaling for high-precision image and video coding are described. A example method of video processing includes performing a conversion between a current block of a video and a bitstream representation of the video according to a rule, wherein the rule specifies that the conversion includes during encoding, skipping applying a forward transform to residual coefficients of the current block prior to including in the bitstream representation, or during decoding, reconstructing residual coefficients of the current block from the bitstream representation without applying an inverse transform, and wherein the rule further specifies that a scale factor is applied to the residual coefficients independent of a size of the current block.

Abstract: A method performed by a decoder for decoding a bitstream comprising a picture parameter set, PPS, and a first set of slices. The method includes obtaining the picture parameter set. The method also includes decoding a syntax element included in the picture parameter set to obtain an indicator value. The decoder is configured such that if the indicator value is set to a first value then the decoder determines that a picture header included in the bitstream comprises a parameter value corresponding to a particular parameter, otherwise the decoder determines that each slice included in the first set of slices comprises a parameter value corresponding to the particular parameter. If the picture header comprises the parameter value corresponding to the particular parameter, then this parameter value is used to decode slice data of each slice included in the first set of slices.

Abstract: In a video coding device, a generation unit generates the prediction error signal from an input image, a transformation unit transforms the prediction error signal depending on a color space selected from among the plurality of color spaces, an adaptive chroma quantization offset derivation unit derives an adaptive chroma quantization offset, which is a parameter for adjusting a quantization parameter of a chroma component, for each of the plurality of color spaces, a quantization unit generates a quantization coefficient image based on the prediction error signal or the transformed prediction error signal by the transformation means, and an inverse quantization unit inverse-quantizes the quantization coefficient image by using the adaptive chroma quantization offset for the selected color space.

Abstract: A vehicular vision system includes a plurality of cameras disposed at a vehicle and capturing image data as the equipped vehicle is driven along a traffic lane of a road. Captured image data is processed to detect a rearward-approaching vehicle and determine a path of travel of the detected rearward-approaching vehicle. Responsive to determining that the path of travel of the detected rearward-approaching vehicle is along a side traffic lane adjacent to the traffic lane along which the equipped vehicle is driven, the vehicular vision system (i) displays at a video display screen video images derived at least from image data captured by the side camera that is at the side portion of the equipped vehicle that is adjacent to the side traffic lane and (ii) overlays a transparent graphic overlay overlaying the displayed video images that is based on the determined path of travel of the detected rearward-approaching vehicle.

Abstract: A compression system trains a compression model for an encoder and decoder. In one embodiment, the compression model includes a machine-learned in-loop flow predictor that generates a flow prediction from previously reconstructed frames. The machine-learned flow predictor is coupled to receive a set of previously reconstructed frames and output a flow prediction for a target frame that is an estimation of the flow for the target frame. In particular, since the flow prediction can be generated by the decoder using the set of previously reconstructed frames, the encoder may transmit a flow delta that indicates a difference between the flow prediction and the actual flow for the target frame, instead of transmitting the flow itself. In this manner, the encoder can transmit a significantly smaller number of bits to the receiver, improving computational efficiency.

Abstract: Provided are systems and methods of using of optical delay lines in RF imagers, e.g., Ultra-wideband (UWB) imagers. In an embodiment, a modulator can be configured to convert radio-frequency signals to optical signal. First and second optical delay lines delay respective first and second optical signals converted by the modulator, and a photodetector can convert the delayed optical signals to at least one electrical signal corresponding to at least one pixel of a radio frequency image. The disclosed systems and methods can also further form a radio-frequency image based on output from the photodetector. In still further embodiments, the photodetector can receive modulated optical signals from an array of optical delays. Also provided are related methods of using the disclosed systems and devices.

Abstract: In various embodiments, a training application trains a machine learning model to preprocess images. In operation, the training application computes a chroma sampling factor based on a downscaling factor and a chroma subsampling ratio. The training application executes a machine learning model that is associated with the chroma sampling factor on data that corresponds to both an image and a first chroma component to generate preprocessed data corresponding to the first chroma component. Based on the preprocessed data, the training application updates at least one parameter of the machine learning model to generate a trained machine learning model that is associated with the first chroma component.

Abstract: A method of decoding encoded information of a point cloud may be performed by at least one processor and comprises: obtaining an encoded bitstream, the encoded bitstream including encoded information of a point cloud including a set of points in a three-dimensional space; and determining a type of partitioning used to encode the information of the point cloud by at least one of parsing signals of at least three binary syntaxes or inferring at least one syntax of the at least three binary syntaxes.

Abstract: An encoder includes circuitry and a memory coupled to the circuitry, wherein the circuitry, in operation, performs a partition process. The partition process includes calculating first values of a set of pixels between a first partition and a second partition in a current block, using a first motion vector for the first partition; calculating second values of the set of pixels, using a second motion vector for the second partition; and calculating third values of the set of pixels by weighting the first values and the second values. When a ratio of a width to a height of the current block is larger than 4 or a ratio of the height to the width of the current block is larger than 4, the circuitry disables the partition process.

Abstract: A merge candidate list is generated, a merge candidate is selected from the merge candidate list as a merge candidate, a bitstream is decoded to derive a motion vector difference, and a corrected merge candidate is derived by adding the motion vector difference to a motion vector of the selected merge candidate for a first prediction without scaling and subtracting the motion vector difference from a motion vector of the selected merge candidate for a second prediction without scaling.

Abstract: Lossy or lossless compression and transmission, comprising the steps of: (i) receiving an input image; (ii) encoding it using an encoder trained neural network, to produce a y latent representation; (iii) encoding the y latent representation using a hyperencoder trained neural network, to produce a z hyperlatent representation; (iv) quantizing the z hyperlatent representation using a predetermined entropy parameter to produce a quantized z hyperlatent representation; (v) entropy encoding the quantized z hyperlatent representation into a first bitstream, using predetermined entropy parameters; (vi) processing the quantized z hyperlatent representation using a hyperdecoder trained neural network to obtain a location entropy parameter ?y, an entropy scale parameter ?y, and a context matrix Ay of the y latent representation; (vii) processing the y latent representation, the location entropy parameter ?y and the context matrix Ay, to obtain quantized latent residuals; (viii) entropy encoding the quantized latent r

Abstract: A method for video processing is provided to include: performing a conversion between a current video block of a video that is a chroma block and a coded representation of the video, wherein, during the conversion, the current video block is constructed based on a first domain and a second domain, and wherein the conversion further includes applying a forward reshaping process and/or an inverse reshaping process to one or more chroma components of the current video block.

Abstract: A method for visual media processing, including performing a conversion between a current chroma video block of visual media data and a bitstream representation of the current chroma video block, wherein, during the conversion, a chroma residual of the current chroma video block is scaled based on a scaling coefficient, wherein the scaling coefficient is derived at least based on luma samples located in predefined positions.

Abstract: A system of eye tracking includes an infrared (IR) source to project IR light to an eye of a user, an IR holographic optical element (HOE) to change an angle of the IR light reflected from the eye of the user, and a sensor to receive the IR light. The system further includes a processor to use the IR light to determine a gaze vector of the user, in one embodiment.

Abstract: A method of video processing includes determining, for a current block, at least one motion candidate list; and performing a conversion between the current block and a bitstream representation of the current block using the at least one motion candidate list, the at least one motion candidate list including at least one motion candidate derived from a set of neighboring blocks including one or more spatial and temporal neighboring blocks.

Abstract: A system includes a processor and a memory. The memory stores instructions executable by the processor to receive an image from a stationary camera. The memory stores instructions to determine location coordinates of an object identified in the image based on location coordinates specified for the image. The memory stores instructions to operate a vehicle based on the object location coordinates.

Abstract: A system for covering an area of interest in a live event is provided. The system may include a masthead, a mast, and a base. The masthead may include a first lens and a second lens. The base may be communicatively coupled to the masthead, and may include base processors configured to determine a position of the system, determine an orientation of the masthead relative to a field of interest, calibrate a first field of view and a second field of view, identify areas of interest, record a location and time stamp of one or more objects, merge a first video and a second video into a merged video, translate a corresponding position of the objects to the merged video, generate a video window based on the translated corresponding position of the objects, and automatically adjust the video window based on changes in the objects.

Abstract: Video compression and decompression techniques are disclosed that provide improved bandwidth control for video compression and decompression systems. In particular, video coding and decoding techniques quantize input video in multiple dimensions. According to these techniques, pixel residuals may be generated from a comparison of an array of input data to an array of prediction data. The pixel residuals may be quantized in a first dimension. After the quantization, the quantized pixel residuals may be transformed to an array of transform coefficients. The transform coefficients may be quantized in a second dimension and entropy coded. Decoding techniques invert these processes. In still other embodiments, multiple quantizers may be provided upstream of the transform stage, either in parallel or in cascade, which provide greater flexibility to video coders to quantize data in different dimensions in an effort to balance the competing interest in compression efficiency and quality of reconstructed video.

Abstract: A method for visual media processing includes performing a conversion between a current video unit of a visual media data and a bitstream representation of the current video unit, wherein the bitstream representation is configured according to a format rule; and wherein the format rule specifies a level of a video segment at which one or more syntax elements indicative of a scaling factor by which a chroma residual is scaled for coding the current video unit is selectively included in the bitstream representation.

Abstract: A vehicular multi-camera vision system includes a plurality of cameras and an electronic control unit (ECU). The cameras are disposed at a vehicle and have respective fields of view exterior of the vehicle. The cameras are operable to capture image data. The cameras are connected with the ECU via respective coaxial cables. Each respective coaxial cable (i) carries captured image data from the respective camera to the ECU, (ii) carries camera control data from the ECU to the respective camera and (iii) connects a DC power supply of the ECU to the respective camera. The camera control data is carried by the respective coaxial cable from the ECU to the respective camera at a control data carrier frequency that is lower than an image data carrier frequency at which image data captured by the respective camera is carried by the respective coaxial cable from the respective camera to the ECU.