Abstract: End-to-end neural image compression using deep reinforcement learning (DRL) is performed by at least one processor and includes encoding an input, generating encoded representations of the input, generating a set of quantization keys using a first neural network, based on a set of previous quantization states, wherein each quantization key in the set of quantization keys and each previous quantization state in the set of previous quantization states correspond to the encoded representations of the input, generating a set of dequantized numbers representing dequantized representations of the encoded representations of the input, based on the set of quantization keys, using a second neural network, and generating a reconstructed output, based on the set of dequantized numbers.

Abstract: A method for image decoding, according to the present invention, includes the following steps: receiving image information including a plurality of filter coefficients; generating a restored block for a current block on the basis of the image information; and applying an adaptive loop filter to the restored block on the basis of the plurality of filter coefficients. According to the present invention, image encoding efficiency may be improved, and complexity may be reduced.

Abstract: Methods and apparatuses for video encoding/decoding are provided. In some examples, an apparatus for video decoding includes receiving circuitry and processing circuitry. The processing circuitry decodes prediction information of a current block in a current picture from a coded video bitstream. The prediction information is indicative of an inter prediction mode. Then, the processing circuitry determines a usage of a first DMVD process on the current block at least partially based on a second DMVD process not being used on the current block. Further, the processing circuitry reconstructs the current block according to the inter prediction mode with the first DMVD process.

Abstract: A better compromise between encoding complexity and achievable rate distortion ratio, and/or to achieve a better rate distortion ratio is achieved by using multitree sub-divisioning not only in order to subdivide a continuous area, namely the sample array, into leaf regions, but using the intermediate regions also to share coding parameters among the corresponding collocated leaf blocks. By this measure, coding procedures performed in tiles—leaf regions—locally, may be associated with coding parameters individually without having to, however, explicitly transmit the whole coding parameters for each leaf region separately. Rather, similarities may effectively exploited by using the multitree subdivision.

Abstract: A depth sensing camera system that comprises one or more fisheye lenses and infrared and/or near-infrared image sensors. In some examples, the image sensors may generate output signals based at least in part on receiving radiation via the fisheye lenses. A depth measurement may be calculated based at least in part on the output signals. For example, these output signals may be provided as input to a depth model, which may determine the depth measurement. In some examples, such a depth model may be integrated into an application-specific integrated circuit and/or may be operated by an application processor.

Abstract: An encoder that encodes a current block in a picture includes circuitry and memory. Using the memory, the circuitry: performs a first transform on a residual signal of the current block using a first transform basis to generate first transform coefficients; and performs a second transform on the first transform coefficients using a second transform basis to generate second transform coefficients and quantizes the second transform coefficients, when the first transform basis is the same as a predetermined transform basis; and quantizes the first transform coefficients without performing the second transform, when the first transform basis is different from the predetermined transform basis.

Abstract: This disclosure describes entropy coding techniques for media data coded using neural-based techniques. A media coder is configured to determine a probability distribution function parameter for a data element of a data stream coded by a neural-based media compression technique, wherein the probability distribution function parameter is a function of a standard deviation of a probability distribution function of the data stream, determine a code vector based on the probability distribution function parameter, and entropy code the data element using the code vector.

Abstract: A method for image decoding performed by a decoding apparatus according to the present disclosure comprises the steps of: deriving an ATMVP candidate for a current block; deriving merge candidates for the current block; deriving prediction samples for the current block on the basis of the merge candidates and the ATMVP candidate; and deriving reconstruction samples for the current block on the basis of the prediction samples, wherein the ATMVP candidate is a spatial neighboring block spatially adjacent to the current block.

Abstract: A better compromise between encoding complexity and achievable rate distortion ratio, and/or to achieve a better rate distortion ratio is achieved by using multitree sub-divisioning not only in order to subdivide a continuous area, namely the sample array, into leaf regions, but using the intermediate regions also to share coding parameters among the corresponding collocated leaf blocks. By this measure, coding procedures performed in tiles—leaf regions—locally, may be associated with coding parameters individually without having to, however, explicitly transmit the whole coding parameters for each leaf region separately. Rather, similarities may effectively exploited by using the multitree subdivision.

Abstract: A method of obtaining quantization parameter (QP) for chrominance components based on QP for luminance component, wherein the method is performed by a decoder or encoder, and wherein the method comprises obtaining a luminance QP for a luma component of a coding unit; obtaining a QP index (QPi) based at least in a part on the luminance QP; and obtaining a chrominance QP (QPc) for a chroma component of the coding unit based on the QP index by using a preset lookup table, wherein the maximum difference between the QP index and the chrominance QP is equal to 3.

Abstract: An image coding method includes: performing context arithmetic coding to consecutively code (i) first information indicating whether or not to perform sample adaptive offset (SAO) processing for a first region of an image and (ii) second information indicating whether or not to use, in the SAO processing for the first region, information on SAO processing for a region other than the first region, the context arithmetic coding being arithmetic coding using a variable probability, the SAO processing being offset processing on a pixel value; and performing bypass arithmetic coding to code other information which is information on the SAO processing for the first region and different from the first information or the second information, after the first information and the second information are coded, the bypass arithmetic coding being arithmetic coding using a fixed probability.

Abstract: An encoder comprising a processor configured to obtain candidate motion vectors (MVs) corresponding to neighboring blocks of a current block, the neighboring blocks neighboring the current block; obtain precisions of the candidate MVs; round the precisions to a target precision based on a rounding scheme; round the candidate MVs based on the target precision; perform pruning of the candidate MVs; generate a candidate list based on the rounding of the candidate MVs and the pruning; select one of the candidate MVs from the candidate list for encoding the current block; and encode an MV candidate index corresponding to the one of the candidate MVs that was selected in a bitstream.

Abstract: A moving picture coding apparatus includes: an inter prediction control unit which determines to code a motion vector using, among candidate predicted motion vectors, a candidate predicted motion vector having the least error with relative to a motion vector derived by motion estimation; a picture type determination unit which generates picture type information; a temporal direction vector calculation unit which derives a candidate predicted motion vector in temporal direct; and a co-located reference direction determination unit which generates, for each picture, a co-located reference direction flag.

Abstract: Various systems and methods for surveillance using a combination of video image capture and passive wireless detection are described. In some cases, the methods include receiving a device identification information from a first wireless access point at a first location and corresponding to a first time, and receiving the device identification from a second wireless access point at a second location and corresponding to a second time. A video from a camera is received, and a travel path is assembled including a portion of the video.

Abstract: Techniques are described herein for processing video data. For instance, a technique can include encoding one or more frames of video data using a video encoder, the video encoder including at least a quantization process; determining an actual bit rate of the encoded one or more frames; predicting an estimated bit rate using an encoder proxy, the encoder proxy including a statistical model for estimating a bit rate of the encoded one or more frames; determining, using the encoder proxy, a gradient of the estimated bit rate; and training the encoder proxy to predict the estimated bit rate based on the actual bit rate, the estimated bit rate, and the gradient.

Abstract: A video encoder is configured to entropy encode an absolute value of a quantization level of a current transform coefficient at position (xC, yC), wherein the absolute value is encoded using context adaptive binary arithmetic encoding of bins of a first binarization. The video coder is further configured, for the current transform coefficient, to encode a bin of the first binarization by using a context which is determined based on a sum of minimum absolute values of transform coefficient quantization levels, based on encoded bins of the first binarization, at one or more transform coefficient positions among (xC+1, yC), (xC+2, yC), (xC+1, yC+1), (xC, yC+1), and (xC, yC+2).

Abstract: Technology to encode/compress and decode/uncompress video streams is described. According to one example, a computer-implemented method includes receiving, by a server and from a first device associated with a first participant of a videoconference involving a second device associated with a second participant, a first compressed video stream and a first decoding model. The first decoding model is trained via machine learning, and the first compressed video stream is compressed using a first encoding model trained via machine learning. The first encoding and decoding models are customized to the first participant. The server is adapted to transmit to the second device, the first decoding model and the first compressed video stream.

Abstract: An encoder calculates an indication to a previous reference picture having temporal identity of zero. The encoder creates a first set of indicators to the previous reference picture, to all reference pictures in a first reference picture set of the previous reference picture, and to all pictures following the previous reference picture in decoding order and precede the current picture in decoding order. The encoder sets a flag for picture order count cycle, when a long term reference picture (LTRP) has least significant bits (LSBs) of a picture order count, for which more than one picture in the first set share same value of the LSBs of picture order count as the LTRP. The decoder obtains LSB of a picture order count for a LTRP in a reference picture set of the current picture. The decoder concludes non-compliant bitstream based on indications provided by the flag.

Abstract: A method of video decoding in a decoder is provided. In the method, a coded video bitstream is received. For a scan position in the transform block, an offset value is determined based on a template magnitude for a template of the scan position. The offset value is constrained based on a first number of context models for each frequency region. For the scan position, a base value is determined based on the first number and the scan position. A context model index is determined based on a sum of the offset value and the base value. A context model is selected from a plurality of context models based on the context model index. A value of a syntax element at the scan position is determined based on the context model. A transform coefficient at the scan position is determined based on the value of the syntax element.

Abstract: According to one embodiment, an image synthesis device for an electronic mirror includes a rear camera which obtains a first image from a first position, and a side camera which obtains a second image of the same direction with a view of the rear camera from a second position. The second image includes a view obstruction. When a part of the first image is connected as a complementary image to the view obstruction of the second image, an image processing device converts an image of the view obstruction into a translucent image, superimposes the complementary image on the translucent image, and obtains a third image which remains an outline of the complementary image.