EP4721405A1 - Image feature pre-processing for reference picture resampling (rpr) decision - Google Patents

Abstract

In one implementation, image features including a down-up PSNR, HOG (Histogram of Oriented Gradients) features and DCT coefficients are extracted, where the down-up PSNR is calculated between the original image and a downscaled (then rescaled to the original resolution) picture, the HOG features are calculated on patches extracted from the original image, and the DCT coefficients are also calculated on the patches. For the DCT coefficients, only a subset at the high frequencies is used. These patches are aggregated and concatenated before feeding to a neural network to predict a QP (Quantization Parameter) switch value. If the QP for the current picture is greater than the predicted QP switch value, then RPR (Reference Picture Resampling) is applied to the current inter picture.

Description

IMAGE FEATURE PRE-PROCESSING FOR REFERENCE PICTURE RESAMPLING (RPR) DECISION

TECHNICAL FIELD

[1] The present embodiments generally relate to a method and an apparatus for reference picture resampling decision in video encoding and decoding.

BACKGROUND

[2] To achieve high compression efficiency, image and video coding schemes usually employ prediction and transform to leverage spatial and temporal redundancy in the video content. Generally, intra or inter prediction is used to exploit the intra or inter picture correlation, then the differences between the original block and the predicted block, often denoted as prediction errors or prediction residuals, are transformed, quantized, and entropy coded. To reconstruct the video, the compressed data are decoded by inverse processes corresponding to the entropy coding, quantization, transform, and prediction.

SUMMARY

[3] According to an embodiment, a method of video encoding is presented, comprising: transforming a plurality of regions of a picture to obtain a plurality sets of transform coefficients, respectively; obtaining a first feature based on the plurality sets of transform coefficients; and predicting a value based on the first feature, wherein the value is used to indicate whether reference picture resampling is to be applied to encode the picture.

[4] According to another embodiment, an apparatus for video encoding is presented, comprising one or more processors, wherein the one or more processors are configured to: transform a plurality of regions of a picture to obtain a plurality sets of transform coefficients, respectively; obtain a first feature based on the plurality sets of transform coefficients; and predict a value based on the first feature, wherein the value is used to indicate whether reference picture resampling is to be applied to encode the picture.

[5] One or more embodiments also provide a computer program comprising instructions which when executed by one or more processors cause the one or more processors to perform the encoding method according to any of the embodiments described herein. One or more of the present embodiments also provide a computer readable storage medium having stored thereon instructions for encoding a video according to the methods described herein. [6] One or more embodiments also provide a computer readable storage medium having stored thereon video data generated according to the methods described above. One or more embodiments also provide a method and apparatus for transmitting or receiving the video data generated according to the methods described herein.

BRIEF DESCRIPTION OF THE DRAWINGS

[7] FIG. 1 illustrates a block diagram of a system within which aspects of the present embodiments may be implemented.

[8] FIG. 2 illustrates a block diagram of an embodiment of a video encoder.

[9] FIG. 3 illustrates a block diagram of an embodiment of a video decoder.

[10] FIG. 4 illustrates resampling for inter-prediction in an encoder.

[11] FIG. 5 illustrates resampling for inter-prediction in a decoder.

[12] FIG. 6 illustrates picture resolution changes with usage of RPR for inter-prediction.

[13] FIG. 7 illustrates RD curves (anchor vs. RPR).

[14] FIG. 8 illustrates a method for RPR mode prediction, according to an embodiment.

[15] FIG. 9 illustrates examples of HOGs.

[16] FIG. 10 illustrates Discrete Cosine Transform (DCT) coefficients.

[17] FIG. 11 illustrates aggregation and concatenation of HOG vectors, DCT coefficients and down-up PSNR value, according to an embodiment.

[18] FIG. 12 illustrates several approaches to select DCT coefficients.

[19] FIG. 13 illustrates a fully connected neural network to predict the QP switch value, according to an embodiment.

DETAILED DESCRIPTION

[20] FIG. 1 illustrates a block diagram of an example of a system in which various aspects and embodiments can be implemented. System 100 may be embodied as a device including the various components described below and is configured to perform one or more of the aspects described in this application. Examples of such devices, include, but are not limited to, various electronic devices such as personal computers, laptop computers, smartphones, tablet computers, digital multimedia set top boxes, digital television receivers, personal video recording systems, connected home appliances, and servers. Elements of system 100, singly or in combination, may be embodied in a single integrated circuit, multiple ICs, and/or discrete components. For example, in at least one embodiment, the processing and encoder/decoder elements of system 100 are distributed across multiple ICs and/or discrete components. In various embodiments, the system 100 is communicatively coupled to other systems, or to other electronic devices, via, for example, a communications bus or through dedicated input and/or output ports. In various embodiments, the system 100 is configured to implement one or more of the aspects described in this application.

[21] The system 100 includes at least one processor 110 configured to execute instructions loaded therein for implementing, for example, the various aspects described in this application. Processor 110 may include embedded memory, input output interface, and various other circuitries as known in the art. The system 100 includes at least one memory 120 (e.g., a volatile memory device, and/or a non-volatile memory device). System 100 includes a storage device 140, which may include non-volatile memory and/or volatile memory, including, but not limited to, EEPROM, ROM, PROM, RAM, DRAM, SRAM, flash, magnetic disk drive, and/or optical disk drive. The storage device 140 may include an internal storage device, an attached storage device, and/or a network accessible storage device, as non-limiting examples.

[22] System 100 includes an encoder/decoder module 130 configured, for example, to process data to provide an encoded video or decoded video, and the encoder/decoder module 130 may include its own processor and memory. The encoder/decoder module 130 represents module(s) that may be included in a device to perform the encoding and/or decoding functions. As is known, a device may include one or both of the encoding and decoding modules. Additionally, encoder/decoder module 130 may be implemented as a separate element of system 100 or may be incorporated within processor 110 as a combination of hardware and software as known to those skilled in the art.

[23] Program code to be loaded onto processor 110 or encoder/decoder 130 to perform the various aspects described in this application may be stored in storage device 140 and subsequently loaded onto memory 120 for execution by processor 110. In accordance with various embodiments, one or more of processor 110, memory 120, storage device 140, and encoder/decoder module 130 may store one or more of various items during the performance of the processes described in this application. Such stored items may include, but are not limited to, the input video, the decoded video or portions of the decoded video, the bitstream, matrices, variables, and intermediate or final results from the processing of equations, formulas, operations, and operational logic.

[24] In several embodiments, memory inside of the processor 110 and/or the encoder/ decoder module 130 is used to store instructions and to provide working memory for processing that is needed during encoding or decoding. In other embodiments, however, a memory external to the processing device (for example, the processing device may be either the processor 110 or the encoder/decoder module 130) is used for one or more of these functions. The external memory may be the memory 120 and/or the storage device 140, for example, a dynamic volatile memory and/or a non-volatile flash memory. In several embodiments, an external non-volatile flash memory is used to store the operating system of a television. In at least one embodiment, a fast external dynamic volatile memory such as a RAM is used as working memory for video coding and decoding operations, such as for MPEG-2, HEVC, or VVC.

[25] The input to the elements of system 100 may be provided through various input devices as indicated in block 105. Such input devices include, but are not limited to, (i) an RF portion that receives an RF signal transmitted, for example, over the air by a broadcaster, (ii) a Composite input terminal, (iii) a USB input terminal, and/or (iv) an HDMI input terminal.

[26] In various embodiments, the input devices of block 105 have associated respective input processing elements as known in the art. For example, the RF portion may be associated with elements suitable for (i) selecting a desired frequency (also referred to as selecting a signal, or band-limiting a signal to a band of frequencies), (ii) down converting the selected signal, (iii) band-limiting again to a narrower band of frequencies to select (for example) a signal frequency band which may be referred to as a channel in certain embodiments, (iv) demodulating the down converted and band-limited signal, (v) performing error correction, and (vi) demultiplexing to select the desired stream of data packets. The RF portion of various embodiments includes one or more elements to perform these functions, for example, frequency selectors, signal selectors, band-limiters, channel selectors, filters, downconverters, demodulators, error correctors, and demultiplexers. The RF portion may include a tuner that performs various of these functions, including, for example, down converting the received signal to a lower frequency (for example, an intermediate frequency or a near-baseband frequency) or to baseband. In one set-top box embodiment, the RF portion and its associated input processing element receives an RF signal transmitted over a wired (for example, cable) medium, and performs frequency selection by filtering, down converting, and filtering again to a desired frequency band. Various embodiments rearrange the order of the above-described (and other) elements, remove some of these elements, and/or add other elements performing similar or different functions. Adding elements may include inserting elements in between existing elements, for example, inserting amplifiers and an analog-to-digital converter. In various embodiments, the RF portion includes an antenna.

[27] Additionally, the USB and/or HDMI terminals may include respective interface processors for connecting system 100 to other electronic devices across USB and/or HDMI connections. It is to be understood that various aspects of input processing, for example, Reed- Solomon error correction, may be implemented, for example, within a separate input processing IC or within processor 110 as necessary. Similarly, aspects of USB or HDMI interface processing may be implemented within separate interface ICs or within processor 110 as necessary. The demodulated, error corrected, and demultiplexed stream is provided to various processing elements, including, for example, processor 110, and encoder/decoder 130 operating in combination with the memory and storage elements to process the datastream as necessary for presentation on an output device.

[28] Various elements of system 100 may be provided within an integrated housing, Within the integrated housing, the various elements may be interconnected and transmit data therebetween using suitable connection arrangement 115, for example, an internal bus as known in the art, including the I2C bus, wiring, and printed circuit boards.

[29] The system 100 includes communication interface 150 that enables communication with other devices via communication channel 190. The communication interface 150 may include, but is not limited to, a transceiver configured to transmit and to receive data over communication channel 190. The communication interface 150 may include, but is not limited to, a modem or network card and the communication channel 190 may be implemented, for example, within a wired and/or a wireless medium.

[30] Data is streamed to the system 100, in various embodiments, using a Wi-Fi network such as IEEE 802. 11. The Wi-Fi signal of these embodiments is received over the communications channel 190 and the communications interface 150 which are adapted for WiFi communications. The communications channel 190 of these embodiments is typically connected to an access point or router that provides access to outside networks including the Internet for allowing streaming applications and other over-the-top communications. Other embodiments provide streamed data to the system 100 using a set-top box that delivers the data over the HDMI connection of the input block 105. Still other embodiments provide streamed data to the system 100 using the RF connection of the input block 105.

[31] The system 100 may provide an output signal to various output devices, including a display 165, speakers 175, and other peripheral devices 185. The other peripheral devices 185 include, in various examples of embodiments, one or more of a stand-alone DVR, a disk player, a stereo system, a lighting system, and other devices that provide a function based on the output of the system 100. In various embodiments, control signals are communicated between the system 100 and the display 165, speakers 175, or other peripheral devices 185 using signaling such as AV. Link, CEC, or other communications protocols that enable device-to-device control with or without user intervention. The output devices may be communicatively coupled to system 100 via dedicated connections through respective interfaces 160, 170, and 180. Alternatively, the output devices may be connected to system 100 using the communications channel 190 via the communications interface 150. The display 165 and speakers 175 may be integrated in a single unit with the other components of system 100 in an electronic device, for example, a television. In various embodiments, the display interface 160 includes a display driver, for example, a timing controller (T Con) chip.

[32] The display 165 and speaker 175 may alternatively be separate from one or more of the other components, for example, if the RF portion of input 105 is part of a separate set-top box. In various embodiments in which the display 165 and speakers 175 are external components, the output signal may be provided via dedicated output connections, including, for example, HDMI ports, USB ports, or COMP outputs.

[33] FIG. 2 illustrates an example video encoder 200, such as a VVC (Versatile Video Coding) encoder. FIG. 2 may also illustrate an encoder in which improvements are made to the VVC standard or an encoder employing technologies similar to VVC.

[34] In the present application, the terms “reconstructed” and “decoded” may be used interchangeably, the terms “encoded” or “coded” may be used interchangeably, and the terms “image,” “picture” and “frame” may be used interchangeably. Usually, but not necessarily, the term “reconstructed” is used at the encoder side while “decoded” is used at the decoder side.

[35] Before being encoded, the video sequence may go through pre-encoding processing (201), for example, applying a color transform to the input color picture (e.g., conversion from RGB 4:4:4 to YCbCr 4:2:0), or performing a remapping of the input picture components in order to get a signal distribution more resilient to compression (for instance using a histogram equalization of one of the color components). Metadata can be associated with the preprocessing, and attached to the bitstream.

[36] In the encoder 200, a picture is encoded by the encoder elements as described below. The picture to be encoded is partitioned (202) and processed in units of, for example, CUs. Each unit is encoded using, for example, either an intra or inter mode. When a unit is encoded in an intra mode, it performs intra prediction (260). In an inter mode, motion estimation (275) and compensation (270) are performed. The encoder decides (205) which one of the intra mode or inter mode to use for encoding the unit, and indicates the intra/inter decision by, for example, a prediction mode flag. After prediction, prediction enhancement (285) is applied to the prediction block. Prediction residuals are calculated, for example, by subtracting (210) the predicted block from the original image block.

[37] The prediction residuals are then transformed (225) and quantized (230). The quantized transform coefficients, as well as motion vectors and other syntax elements, are entropy coded (245) to output a bitstream. The encoder can skip the transform and apply quantization directly to the non-transformed residual signal. The encoder can bypass both transform and quantization, i.e., the residual is coded directly without the application of the transform or quantization processes.

[38] The encoder decodes an encoded block to provide a reference for further predictions. The quantized transform coefficients are de-quantized (240) and inverse transformed (250) to decode prediction residuals. Combining (255) the decoded prediction residuals and the predicted block, an image block is reconstructed. In-loop filters (265) are applied to the reconstructed picture to perform, for example, deblocking/SAO (Sample Adaptive Offset) filtering to reduce encoding artifacts. The filtered image is stored at a reference picture buffer (280).

[39] FIG. 3 illustrates a block diagram of an example video decoder 300. In the decoder 300, a bitstream is decoded by the decoder elements as described below. Video decoder 300 generally performs a decoding pass reciprocal to the encoding pass as described in FIG. 2. The encoder 200 also generally performs video decoding as part of encoding video data.

[40] In particular, the input of the decoder includes a video bitstream, which can be generated by video encoder 200. The bitstream is first entropy decoded (330) to obtain transform coefficients, motion vectors, and other coded information. The picture partition information indicates how the picture is partitioned. The decoder may therefore divide (335) the picture according to the decoded picture partitioning information. The transform coefficients are de-quantized (340) and inverse transformed (350) to decode the prediction residuals. Combining (355) the decoded prediction residuals and the predicted block, an image block is reconstructed. The predicted block can be obtained (370) from intra prediction (360) or motion-compensated prediction (i.e., inter prediction) (375). After prediction, prediction enhancement (390) is applied to the prediction block. In-loop filters (365) are applied to the reconstructed image. The filtered image is stored at a reference picture buffer (380).

[41] The decoded picture can further go through post-decoding processing (385), for example, an inverse color transform (e.g., conversion from YCbCr 4:2:0 to RGB 4:4:4) or an inverse remapping performing the inverse of the remapping process performed in the preencoding processing (201). The post-decoding processing can use metadata derived in the preencoding processing and signaled in the bitstream.

[42] Reference Picture Resampling (RPR)

[43] In the Versatile Video Coding (VVC)/H.266 standard, the picture-based re-scaling feature for video coding is named Reference Picture Resampling (RPR). Given an original video sequence composed of pictures of size (width x height), the encoder may choose for each frame which resolution (picture size) to use for coding the frame. Different picture parameter sets (PPS) are coded in the bit-stream with the possible sizes of the pictures and the slice/picture header indicates which PPS to use to decode the current video coding layer (VCL) network abstraction layer (NAL) unit.

[44] FIG. 4 is a functional block diagram of an example encoder (400) using RPR. The encoder (400) includes a down-sampler (440), a core encoder (410), a re-sampler (430, a component of motion compensator 450), and a decoded picture buffer (DPB, 420). The encoder (400) is employed to encode an original video into a coded video bitstream. In an aspect, the core encoder (410) represents the encoder (200) of FIG. 2, except for the reference picture buffer (280) and the motion compensator (270) that are shown in FIG. 4 as decoded picture buffer (420) and motion compensator (450), respectively.

[45] In the example of FIG. 4, for each input frame of the original video, the encoder (400) can select whether to encode the frame at the original frame size or at a reduced frame size. If the encoder selects to encode at a reduced frame size, the input frame is down-sampled, by the down-sampler (440), before the frame is encoded by the core encoder (410). For example, the frame may be sub-sampled (440) by a resolution rate of 1/2, resulting in frame size of width W /2 and of height H /2. The decision whether to encode a frame at the original frame size or at a reduced frame size can be done by the encoder (400), for example, by comparing coding results of the input frame at various resolutions or based on spatial and temporal activity related to the coding of input frames of the original video. Consequently, when image content from a currently encoded frame and a corresponding content from a reference frame (needed for interprediction of the content from the currently encoded frame) are not at the same resolution, a resampling operation is required. In VVC, the resampling operation and the motion compensation are combined into one single step. This can be done by the re-sampler (430) that can be configured to change the resolution of image content from reference frames to a resolution that matches the resolution of content from the current frame.

[46] FIG. 5 is a functional block diagram of an example decoder (500) using RPR. The decoder (500) includes a core decoder (510), a re-sampler (530, a component of motion compensator 550), a decoded picture buffer (520), and an up-sampler (540). The decoder (500) is employed to decode the original video from a given coded video bitstream, resulting in the output video. In an aspect, the core decoder (510) represents the decoder (300) of FIG. 3, except for the reference picture buffer (380) and the motion compensator (375) that are shown in FIG. 5 as decoded picture buffer (520) and motion compensator (550), respectively.

[47] In the example of FIG. 5, image content from one or more frames (stored in the decoded picture buffer 520) that is needed as reference to inter-predict image content from a currently decoded frame is re-sampled, by the re-sampler (530), into a resolution that matches the resolution of the currently decoded image content. The video frames that are decoded in a reduced resolution are up-sampled to the resolution of the original video, by the up-sampler (540), resulting in the output video. Note that the down-sampler (440) of FIG. 4 (e.g., employed in a pre-processing stage) and the up-sampler (540) of FIG. 5 (e.g., employed in a post-processing stage) are not specified by the VCC standard.

[48] For each frame, the encoder chooses whether to encode at the original or down-sized resolution (e.g., picture width/height divided by 2). The choice can be made with two-pass encoding or by considering spatial and temporal activity in the original pictures for example. Consequently, the decoded picture buffer (DPB, 420, 520) can contain pictures with different sizes than the current picture size.

[49] In case one reference picture in the DPB has a size different from the current picture, the resampling (430, 530) (up-scaling or down-scaling) of the reference block to build the prediction block is made implicitly during the motion compensation process (270, 375).

[50] FIG. 6 illustrates the application of RPR in a coded stream with successive pictures having different picture resolutions. This feature is for instance useful for adaptive streaming, to adapt the bitrate to the network constraints. Even if it is not normatively specified, the down- sampled pictures, that are not at the resolution indicated in the HLS, need to be resampled to show all decoded pictures on the end-display at a same (display or target) resolution, as illustrated in FIG. 5 (540).

[51] Evaluation of video coding performance

[52] Rate-distortion and complexity are two criteria that are usually used to compare video codec performance. Rate-distortion measures the compression efficiency and gives the relationship between the bitrate (Eq. 1) and the quality of the reconstructed video. Peak Signal to Noise Ratio (PSNR) (Eq. 2) is often used to evaluate the quality, where I(i,j) and K(i,j) represent pixel at (i, j) position in the original image I and pixel at (i, j) position in the reconstructed image K, respectively.

[53] Generally, a tradeoff between the bitrate and the distortion is controlled by a quantization parameter (QP) input. Video codec performance is evaluated by performing the bitrate and PSNR measurement at several QP values, constituting the rate distortion curve (RD- curve).

[54] Traditionally, if one wants to compare the performance of two versions of the same video encoder for a range of QPs, Bjontegaard delta rate (BD-rate) measurement is used to measure the average bitrate and the quality difference between RD-curve of each version of the encoder, as shown in an example in FIG. 7.

[55] In the following we denote by “RPR mode” the resizing (down-sampling) of the frame. In case of RPR vs. non-RPR encoder version, use of RPR improves the BD-rate usually for high QP values. It can be determined at which location the RD curves cross as shown in FIG. 7. Notice that the crossing point, also known as a QP switch, is video content dependent. It does not have the same value and can be different from frame to frame or from group of frames to another. The QP switch could be used to decide whether to apply or not the RPR mode, in this example based on a PSNR performance metric. Note that VVC standard does not present io a method describing when to enable RPR.

[56] One option is to let the RDO (Rate Distortion Optimization) in the encoder to decide when to apply RPR mode. But this requires at least two passes of encoding, one for encoding a full resolution image and the second for down-scaled resolution (one pass for each possible ratio actually). This option has the drawback of putting additional complexity at the encoder in terms of encoding time, and in general that is what one would like to avoid.

[57] Another option, presented in this document, is to train a neural network predictor to predict the QP switch related to the image content and then decide whether to apply RPR or not. In general, the present document introduces a method to pre-process image features, detect and estimate when to enable the RPR mode at the encoder, in order to achieve better coding efficiency. One embodiment is to perform RPR decision relying on the complexity of the video frames and the quantization parameter (QP).

[58] In one embodiment, RPR mode decision is predicted by a fully connected neural network relying on features extracted from video images. A training step is performed to train the neural network weights parameters, by processing a large number of labeled samples that have been prepared offline. Samples here correspond to features (Down-Up PSNR, HOG, DCT) extracted from images of video sequences and their labels (QP switch values). Backpropagation and gradient descent are two algorithms that are usually used to tune the neural network weights in order to minimize the error between the predicted value and the ground-truth label.

[59] The RPR mode prediction according to this embodiment is depicted in FIG. 8. In particular, features are extracted (810) and pre-processed (820). QP switch points can then be predicted (830), and applied to RPR decision (840). In the following, each module is described in further detail.

[60] Feature extraction (810)

[61] The following three types of features can be used to predict the QP switch value for an input video frame.

■ Down-up PSNR: calculates a PSNR between the original image and a downscaled image, e.g., by factor 4 and rescaled to the original resolution. This feature provides information on the quality of the image reconstruction. ■ HOG features: Calculates a Histogram of Oriented Gradients (920) on patches extracted from the original image (910), as illustrated in FIG. 9. This feature counts the occurrence of gradient orientation (930), and usually are calculated on a dense grid (915) of uniform spaced patches obtained from an image. The HOG feature provides information on the distribution of the edge directions.

■ DCT: Discrete Cosine Transform converts image pixels from spatial domain to frequency domain. For example, the image is divided into patches of, for example, 8 by 8 pixels and then DCT is applied on each patch to obtain DCT coefficients as shown in an example in FIG. 10. High frequency coefficients are on bottom-right while low frequency coefficients on top-left. DCT and IDCT are given by equations (3), (4) and (5). This feature provides valuable information on frequency distribution in the original image. f — = n

C(m) = V2 ’" ^{- U} (5) ll, m = 1,2, ...7

[62] Feature pre-processing (820)

[63] During feature pre-processing, global features (e.g., down-up PSNR) and/or local features (e.g., DCT, HOG) are reshaped, aggregated and concatenated to get one global feature vector that feeds a Neural Network input. It acts as a dimensionality reduction. Other global and local features can be used. In one example, luminance average and standard deviation can be used. In another example, Scale-Invariant Feature Transform (SIFT) and Speeded Up Robust Features (SURF) can be used. The feature pre-processing stage has a two-fold objective: on one hand, it is to keep a relevant image descriptor into a small vector size, and on the other hand, it maintains the neural network complexity reasonable in terms of trainable parameters and inference complexity.

[64] FIG. 11 illustrates the aggregation and concatenation of HOG vectors, DCT coefficients and down-up PSNR value, according to this embodiment. Let us take an example of an image of size 32* 32 pixels. After dividing the image in 8 * 8 pixels patches (16 patches), DCT coefficients (1111) and HOG (1121) are calculated on each patch, and reshaped (1110, 1120) along z axis, resulting in 16 * 8 * 8 DCT coefficients (1110) and 16 histograms of oriented gradients (1120).

[65] The DCT aggregation process (1150) consists of averaging or taking the median value per frequency channel (1112) along z axis, ending up with 8 * 8 DCT coefficients values. The high frequency coefficients are the most relevant values for QP switch prediction. Several approaches can be considered to select them and some examples of DCT coefficients are given in FIG. 12. The selected DCT coefficients values are then re-arranged in ID vector (known as flatten/reshaping operation).

[66] In case of HOG aggregation (1160), average or median value is computed along z axis for each bin (1122) of oriented gradient getting ID vector of 9 bins. Note that the number of bins can be set to another value. The down-up PSNR value is calculated (1170) between the original image and a downscaled image (1130), e.g., by factor 4 and rescaled (1140) to the original resolution. Finally, the feature pre-processing module outputs (1180) an ID concatenated vector composed of the down-up PSNR value, ID HOG vector and ID DCT coefficients vector.

[67] QP switch prediction (830)

[68] The concatenated vector resulting from the feature pre-processing module is fed into a fully connected neural network, e.g., as illustrated in FIG. 13, where the number of neurons (1320) per layer and the number of layers (1310) can be adjusted during the training phase in order to get a good performance of the predictor in terms of precision and generalization to new images. Other predictors from the machine learning domain can be used like Support Vectors Regressor (SVR) or Decision Tree Regression. A QP switch value is then predicted (1330) and used to provide an RPR decision.

[69] In the above, the input to the neural network contains the down-up PSNR value, ID HOG vector and ID DCT coefficients vector. Alternatively, the neural network may only take a subset as the input to the neural network for feature pre-processing. For example, the neural network input can be based on just the ID DCT coefficient vector. As DCT provides a compact representation of an image in the spectral domain, it includes useful information on the presence of high/low frequency that impact the prediction. Further, the DCT feature can be concatenated with a second feature based on gradient directions in order to predict the QP switch. This second feature can be complementary to the first one giving a global information on the image gradient orientation. In addition, the down-up PSNR feature can be concatenated to one or both of the other two features to provide an indication of the quality of the reconstruction after downscaling and upscaling the input image.

[70] RPR decision (840)

[71] A simple comparison between the current quantization parameter QP and predicted QP switch is performed to decide whether to apply RPR or not on the current frame. When the current QP is smaller than the predicted QP switch, RPR is not applied; otherwise, RPR is applied.

[72] Various methods are described herein, and each of the methods comprises one or more steps or actions for achieving the described method. Unless a specific order of steps or actions is required for proper operation of the method, the order and/or use of specific steps and/or actions may be modified or combined. Additionally, terms such as “first”, “second”, etc. may be used in various embodiments to modify an element, component, step, operation, etc., for example, a “first decoding” and a “second decoding”. Use of such terms does not imply an ordering to the modified operations unless specifically required. So, in this example, the first decoding need not be performed before the second decoding, and may occur, for example, before, during, or in an overlapping time period with the second decoding.

[73] Various methods and other aspects described in this application can be used to modify modules, for example, the motion compensation module (270) of a video encoder 200 as shown in FIG. 2. Moreover, the present aspects are not limited to ECM and VVC, and can be applied, for example, to other standards and recommendations, and extensions of any such standards and recommendations. Unless indicated otherwise, or technically precluded, the aspects described in this application can be used individually or in combination.

[74] Various numeric values are used in the present application. The specific values are for example purposes and the aspects described are not limited to these specific values.

[75] Various implementations involve decoding. “Decoding,” as used in this application, may encompass all or part of the processes performed, for example, on a received encoded sequence in order to produce a final output suitable for display. In various embodiments, such processes include one or more of the processes typically performed by a decoder, for example, entropy decoding, inverse quantization, inverse transformation, and differential decoding. Whether the phrase “decoding process” is intended to refer specifically to a subset of operations or generally to the broader decoding process will be clear based on the context of the specific descriptions and is believed to be well understood by those skilled in the art.

[76] Various implementations involve encoding. In an analogous way to the above discussion about “decoding”, “encoding” as used in this application may encompass all or part of the processes performed, for example, on an input video sequence in order to produce an encoded bitstream.

[77] Note that the syntax elements as used herein are descriptive terms. As such, they do not preclude the use of other syntax element names.

[78] The implementations and aspects described herein may be implemented in, for example, a method or a process, an apparatus, a software program, a data stream, or a signal. Even if only discussed in the context of a single form of implementation (for example, discussed only as a method), the implementation of features discussed may also be implemented in other forms (for example, an apparatus or program). An apparatus may be implemented in, for example, appropriate hardware, software, and firmware. The methods may be implemented in, for example, an apparatus, for example, a processor, which refers to processing devices in general, including, for example, a computer, a microprocessor, an integrated circuit, or a programmable logic device. Processors also include communication devices, for example, computers, cell phones, portable/personal digital assistants (“PDAs”), and other devices that facilitate communication of information between end-users.

[79] Reference to “one embodiment” or “an embodiment” or “one implementation” or “an implementation”, as well as other variations thereof, means that a particular feature, structure, characteristic, and so forth described in connection with the embodiment is included in at least one embodiment. Thus, the appearances of the phrase “in one embodiment” or “in an embodiment” or “in one implementation” or “in an implementation”, as well any other variations, appearing in various places throughout this application are not necessarily all referring to the same embodiment.

[80] Additionally, this application may refer to “determining” various pieces of information. Determining the information may include one or more of, for example, estimating the information, calculating the information, predicting the information, or retrieving the information from memory.

[81] Further, this application may refer to “accessing” various pieces of information. Accessing the information may include one or more of, for example, receiving the information, retrieving the information (for example, from memory), storing the information, moving the information, copying the information, calculating the information, determining the information, predicting the information, or estimating the information.

[82] Additionally, this application may refer to “receiving” various pieces of information. Receiving is, as with “accessing”, intended to be a broad term. Receiving the information may include one or more of, for example, accessing the information, or retrieving the information (for example, from memory). Further, “receiving” is typically involved, in one way or another, during operations, for example, storing the information, processing the information, transmitting the information, moving the information, copying the information, erasing the information, calculating the information, determining the information, predicting the information, or estimating the information.

[83] It is to be appreciated that the use of any of the following “and/or”, and “at least one of’, for example, in the cases of “A/B”, “A and/or B” and “at least one of A and B”, is intended to encompass the selection of the first listed option (A) only, or the selection of the second listed option (B) only, or the selection of both options (A and B). As a further example, in the cases of “A, B, and/or C” and “at least one of A, B, and C”, such phrasing is intended to encompass the selection of the first listed option (A) only, or the selection of the second listed option (B) only, or the selection of the third listed option (C) only, or the selection of the first and the second listed options (A and B) only, or the selection of the first and third listed options (A and C) only, or the selection of the second and third listed options (B and C) only, or the selection of all three options (A and B and C). This may be extended, as is clear to one of ordinary skill in this and related arts, for as many items as are listed.

[84] Also, as used herein, the word “signal” refers to, among other things, indicating something to a corresponding decoder. For example, in certain embodiments the encoder signals a quantization matrix for de-quantization. In this way, in an embodiment the same parameter is used at both the encoder side and the decoder side. Thus, for example, an encoder can transmit (explicit signaling) a particular parameter to the decoder so that the decoder can use the same particular parameter. Conversely, if the decoder already has the particular parameter as well as others, then signaling can be used without transmitting (implicit signaling) to simply allow the decoder to know and select the particular parameter. By avoiding transmission of any actual functions, a bit savings is realized in various embodiments. It is to be appreciated that signaling can be accomplished in a variety of ways. For example, one or more syntax elements, flags, and so forth are used to signal information to a corresponding decoder in various embodiments. While the preceding relates to the verb form of the word “signal”, the word “signal” can also be used herein as a noun.

[85] As will be evident to one of ordinary skill in the art, implementations may produce a variety of signals formatted to carry information that may be, for example, stored or transmitted. The information may include, for example, instructions for performing a method, or data produced by one of the described implementations. For example, a signal may be formatted to carry the bitstream of a described embodiment. Such a signal may be formatted, for example, as an electromagnetic wave (for example, using a radio frequency portion of spectrum) or as a baseband signal. The formatting may include, for example, encoding a data stream and modulating a carrier with the encoded data stream. The information that the signal carries may be, for example, analog or digital information. The signal may be transmitted over a variety of different wired or wireless links, as is known. The signal may be stored on a processor-readable medium.

Claims

1. A method of video encoding, comprising: transforming a plurality of regions of a picture to obtain a plurality sets of transform coefficients, respectively; obtaining a first feature based on the plurality sets of transform coefficients; and predicting a value based on the first feature, wherein the value is used to indicate whether reference picture resampling is to be applied to encode the picture.

2. An apparatus for video encoding, comprising one or more processors, wherein the one or more processors are configured to: transform a plurality of regions of a picture to obtain a plurality sets of transform coefficients, respectively; obtain a first feature based on the plurality sets of transform coefficients; and predict a value based on the first feature, wherein the value is used to indicate whether reference picture resampling is to be applied to encode the picture.

3. The method of claim 1, or the apparatus of claim 2, wherein the first feature is based on a median value or an average value of transform coefficients, at each frequency, of the plurality sets of transform coefficients.

4. The method of claim 3, or the apparatus of claim 3, wherein the first feature is based on a high frequency part of transform coefficients of the plurality sets of transform coefficients.

5. The method of any one of claims 1, 3 and 4, further comprising, or the apparatus of any one of claims 2-4, wherein the one or more processors are configured to perform: obtaining a second feature indicating a distribution of gradient directions of the plurality of regions of the picture, wherein the value is predicted further based on the second feature.

6. The method of any one of claims 1 and 3-5, or the apparatus of any one of claims 2- 5, wherein a respective histogram of oriented gradients is obtained for each region in the picture, and wherein the second feature is based on a median value or an average value of histogram values, at each bin position, of histograms of oriented gradients.

7. The method of any one of claims 1 and 3-6, further comprising, or the apparatus of any one of claims 2-6, wherein the one or more processors are further configured to perform: downsampling the picture and upsampling the downsampled picture to an original resolution of the picture to form a downsampled and then upsampled picture; and obtaining a third feature indicative of a difference between the picture and the downsampled and then upsampled picture, wherein the value is predicted further based on the third feature.

8. The method of claim 7, or the apparatus of claim 7, wherein the third feature corresponds to a PSNR (Peak Signal to Noise Ratio).

9. The method of any one of claims 1 and 3-8, or the apparatus of any one of claims 2- 8, wherein the value is predicted based on machine learning.

10. The method of claim 8 or 9, or the apparatus of claim 8 or 9, wherein the first, second and third features are concatenated to form an input for a neural network.

11. The method of any one of claims 1 and 3-10, or the apparatus of any one of claims 2-10, wherein reference picture resampling is enabled responsive to a quantization parameter for encoding the picture being greater than the value.

12. The method of any one of claims 1 and 3-11, or the apparatus of any one of claims 2-11, wherein reference picture resampling is disabled responsive to a quantization parameter for encoding the picture being smaller than the value.

13. The method of any one of claims 1 and 3-12, or the apparatus of any one of claims 2-12, wherein the predicting a value is performed by a neural network.

14. A signal comprising video data, formed by performing the method of any one of claims 1 and 3-13.

15. A computer readable storage medium having stored thereon instructions for video encoding according to the method of any one of claims 1 and 3-13.