Rate-distortion optimization for high dynamic range video coding

HDR video quality assessment: Perceptual evaluation of compressed HDR video

The new generation of organic light emitting diode display is designed to enable the high dynamic range (HDR), going beyond the standard dynamic range (SDR) supported by the traditional display devices. However, a large quantity of videos are still of SDR format. Further, most pre-existing videos are compressed at varying degrees for minimizing the storage and traffic flow demands. To enable movie-going experience on new generation devices, converting the compressed SDR videos to the HDR format (i.e., compressed-SDR to HDR conversion) is in great demands. The key challenge with this new problem is how to solve the intrinsic many-to-many mapping issue. However, without constraining the solution space or simply imitating the inverse camera imaging pipeline in stages, existing SDR-to-HDR methods can not formulate the HDR video generation process explicitly. Besides, they ignore the fact that videos are often compressed. To address these challenges, in this work we propose a novel imaging knowledge-inspired parallel networks (termed as KPNet) for compressed-SDR to HDR (CSDR-to-HDR) video reconstruction. KPNet has two key designs: Knowledge-Inspired Block (KIB) and Information Fusion Module (IFM). Concretely, mathematically formulated using some priors with compressed videos, our conversion from a CSDR-to-HDR video reconstruction is conceptually divided into four synergistic parts: reducing compression artifacts, recovering missing details, adjusting imaging parameters, and reducing image noise. We approximate this process by a compact KIB. To capture richer details, we learn HDR representations with a set of KIBs connected in parallel and fused with the IFM. Extensive evaluations show that our KPNet achieves superior performance over the state-of-the-art methods.

High dynamic range (HDR) video compression has until now been approached by using the high profile of existing state-of-the-art H.264/AVC (Advanced Video Coding) codec or by separately encoding low dynamic range (LDR) video and the residue resulted from the estimation of HDR video from LDR video. Although the latter approach has a distinctive advantage of providing backward compatibility to 8-bit LDR displays, the superiority of one approach to the other in terms of the rate distortion trade-off has not been verified yet. In this paper, we first give a detailed overview of the methods in these two approaches. Then, we experimentally compare two approaches with respect to different objective and perceptual metrics, such as HDR mean square error (HDR MSE), perceptually uniform peak signal to noise ratio (PU PSNR) and HDR visible difference predictor (HDR VDP). We first conclude that the optimized methods for backward compatibility to 8-bit LDR displays are superior to the method designed for high profile encoder both for 8-bit and 12-bit mappings in terms of all metrics. Second, using higher bit-depths with a high profile encoder is giving better rate-distortion performances than employing an 8-bit mapping with an 8-bit encoder for the same method, in particular when the dynamic range of the video sequence is high. Third, rather than encoding of the residue signal in backward compatible methods, changing the quantization step size of the LDR layer encoder would be sufficient to achieve a required quality. In other words, the quality of tone mapping is more important than residue encoding for the performance of HDR image and video coding.

There is a rapid increase in the amount of High Dynamic Range (HDR) display devices that are able to offer significantly better viewing experience with higher contrast and richer colors. However, there is a considerable shortage of content that is able to offer the much-sought-after HDR experience. One approach is to capture new content using specialized HDR-cameras, which are typically fewer and costlier than the conventional cameras. Another approach is to convert existing Standard Dynamic Range (SDR) content into HDR artificially, such that, it can reasonably mimic the viewing experience that a true HDR content would have offered. With the latter approach, we propose a Gaussian Process Regression (GPR) based machine learning method for estimating HDR content from their SDR counterparts. GPR is known as a powerful technique for estimating continuous real-valued functions. Given a set of training SDR-HDR image pairs, our proposed method is able to estimate the reverse tone mapping function that is used to convert SDR signal into its HDR equivalent. Preliminary experimental results indicate that our approach produces visually pleasing HDR images.

The final consumer of videos is mostly human. Therefore, if videos can be compressed by fully utilizing the perception characteristics of human visual systems (HVS), the bitrates of the compressed videos can be significantly reduced with subjective visual quality degradation as little as possible. Based on this, we newly propose a learning-based Just Noticeable Distortion (JND)-directed preprocessing scheme for perceptual video compression, especially for 10-bit High Dynamic Range (HDR) videos, which is called the HDR-JNDNet. Our HDR-JNDNet effectively suppresses the perceptual redundancy of 10-bit HDR video signals so that the compression efficiency can be significantly enhanced for the HEVC main10 profile encoder. To our best knowledge, our work is the first approach to training a CNN-based model to directly generate the JND-directed suppressed frames of 10-bit HDR video with the negligible perceptual quality difference between the decoded frames for the original HDR video input with and without the preprocessing by our HDR-JNDNet. Via intensive experiments, when the HDR-JNDNet is applied as preprocessing for the HDR video input before compression, it allows to remarkably save the required bitrates up to the maximum (average) 40.66% (18.37%) for 4K-UHD/HDR test videos, with little subjective video quality degradation without increasing the computational complexity.

High Dynamic Range (HDR) image and video technology aims at conveying the full range of perceptible shadow and highlight details with sufficient precision. HDR is regarded by many experts as the next evolution in digital media. However, industrial broadcasters have concerns regarding the bandwidth overhead that this new technology entails. While many consider that broadcasting HDR content would increase bandwidth requirements by around 20%, this number is based on studies where, in addition to the SDR main stream, HDR-related side information is conveyed. A recent subjective evaluation reported that encoding HDR video content in a single layer might require less bandwidth than its associated SDR version. Similar results were discussed in the MPEG ad-hoc group on High Dynamic Range and Wide Color Gamut. In this article, we explain how having more information can result in lower bandwidth requirements. To this end, we describe several limitations of the human vision system that, when exploited, optimize the HDR distribution pipeline for a human observer. Our theoretical assumption about the higher efficiency of HDR is backed up by a statistical analysis of pixel distribution in real images. The Spatial Index objective metric also reconfirms our assumption.

The migration from High Definition (HD) TV to Ultra High Definition (UHD) is already underway. In addition to an increase of picture spatial resolution, UHD will bring more color and higher contrast by introducing Wide Color Gamut (WCG) and High Dynamic Range (HDR) video. As both Standard Dynamic Range (SDR) and HDR devices will coexist in the ecosystem, the transition from Standard Dynamic Range (SDR) to HDR will require distribution solutions supporting some level of backward compatibility. This paper presents a new HDR content distribution scheme, named SL-HDR1, using a single layer codec design and providing SDR compatibility. The solution is based on a pre-encoding HDR-to-SDR conversion, generating a backward compatible SDR video, with side dynamic metadata. The resulting SDR video is then compressed, distributed and decoded using standard-compliant decoders (e.g. HEVC Main 10 compliant). The decoded SDR video can be directly rendered on SDR displays without adaptation. Dynamic metadata of limited size are generated by the pre-processing and used to reconstruct the HDR signal from the decoded SDR video, using a post-processing that is the functional inverse of the pre-processing. Both HDR quality and artistic intent are preserved. Pre- and post-processing are applied independently per picture, do not involve any inter-pixel dependency, and are codec agnostic. Compression performance, and SDR quality are shown to be solidly improved compared to the non-backward and backward-compatible approaches, respectively using the Perceptual Quantization (PQ) and Hybrid Log Gamma (HLG) Opto-Electronic Transfer Functions (OETF).

Currently, high dynamic range (HDR) videos with high resolution (HR) have become popular due to the display and the rendered technological advancements. However, making ultra-high definition (UHD) with HDR videos is expensive. The legacy low-resolution (LR) standard dynamic range (SDR) format is still largely used in practice. It is necessary to search for a solution to transform LR SDR videos into UHD HDR format. In this paper, we consider joint super resolution and learning inverse tone mapping an issue of high-frequency reconstruction and local contrast enhancement, and we propose an architecture based on a generative adversarial network to apply joint SR-ITM learning. Specifically, we include the residual ResNeXt block (RRXB) as a basic module to better capture high-frequency textures and adopt YUV interpolation to achieve local contrast enhancement. By adopting a generative adversarial network as a pivotal training mechanism, our designs show advantages in both integration and performance. Our code is now available on GitHub: SR-ITM-GAN.

The emergence of the new generation video coding standard, Versatile Video Coding (VVC), has brought along novel features rendering the new standard more efficient and flexible than its predecessors. Aside from efficient compression of 8 k or higher camera-captured content, VVC also supports a wide range of applications, including computer-generated content, high dynamic range (HDR) content, multilayer and multi-view coding, video region extraction, as well as 360° video. One of the newly introduced coding tools in VVC, offering extraction and independent coding of rectangular sub-areas within a frame, is called Subpicture. In this work, we turn our attention to frame partitioning using Subpictures in VVC, and more particularly, a content-aware partitioning is considered. To achieve that, we make use of image segmentation algorithms and properly modify them to operate on a per Coding Tree Unit (CTU) basis in order to render them compliant with the standard’s restrictions. Additionally, since subpicture boundaries need to comply with slice boundaries, we propose two methods for properly partitioning a frame using tiles/slices aiming to avoid over-partitioning of a frame. The proposed algorithms are evaluated regarding both compression efficiency and image segmentation effectiveness. Our evaluation results indicate that the proposed partitioning schemes have a negligible impact on compression efficiency and video quality

Rate-distortion optimized layered coding of high dynamic range videos

As high dynamic range video is gaining popularity, video coding solutions able to efficiently provide both low and high dynamic range video, notably with a single bitstream, are increasingly important. While simulcasting can provide both dynamic range videos at the cost of some compression efficiency penalty, bit-depth scalable video coding can provide a better trade-off between compression efficiency, adaptation flexibility and computational complexity. Considering the widespread use of H.264/AVC video, this paper proposes a H.264/AVC backward compatible bit-depth scalable video coding solution offering a low dynamic range base layer and two high dynamic range enhancement layers with different qualities, at low complexity. Experimental results show that the proposed solution has an acceptable rate-distortion performance penalty regarding the HDR H.264/AVC single-layer coding solution.

A computationally fast tone mapping operator (TMO) that can quickly adapt to a wide spectrum of high dynamic range (HDR) content is quintessential for visualization on varied low dynamic range (LDR) output devices such as movie screens or standard displays. Existing TMOs can successfully tone-map only a limited number of HDR content and require an extensive parameter tuning to yield the best subjective-quality tone-mapped output. In this paper, we address this problem by proposing a fast, parameter-free and scene-adaptable deep tone mapping operator (DeepTMO) that yields a high-resolution and high-subjective quality tone mapped output. Based on conditional generative adversarial network (cGAN), DeepTMO not only learns to adapt to vast scenic-content (e.g., outdoor, indoor, human, structures, etc.) but also tackles the HDR related scene-specific challenges such as contrast and brightness, while preserving the fine-grained details. We explore 4 possible combinations of Generator-Discriminator architectural designs to specifically address some prominent issues in HDR related deep-learning frameworks like blurring, tiling patterns and saturation artifacts. By exploring different influences of scales, loss-functions and normalization layers under a cGAN setting, we conclude with adopting a multi-scale model for our task. To further leverage on the large-scale availability of unlabeled HDR data, we train our network by generating targets using an objective HDR quality metric, namely Tone Mapping Image Quality Index (TMQI). We demonstrate results both quantitatively and qualitatively, and showcase that our DeepTMO generates high-resolution, high-quality output images over a large spectrum of real-world scenes. Finally, we evaluate the perceived quality of our results by conducting a pair-wise subjective study which confirms the versatility of our method.

In this paper, we propose a bit-depth scalable lossless coding method for high dynamic range (HDR) images based on a reversible logarithmic mapping. HDR images are generally expressed as floating-point data, such as in the OpenEXR or RGBE formats. Our bit-depth scalable coding approach outputs base layer data and enhancement layer data. It can reconstruct the low dynamic range (LDR) image from the base layer data and reconstructs the HDR image by adding the enhancement layer data. Most previous two-layer methods have focused on the lossy coding of HDR images. Unfortunately, the extension of previous lossy methods to lossless coding does not significantly compress the enhancement layer data. This is because the bit depth becomes very large, especially for HDR images in floating-point data format. To tackle this problem, we apply a reversible logarithmic mapping to the input HDR data. Moreover, we introduce a format conversion to avoid any degradation in the quality of the reconstructed LDR image. The proposed method is effective for both OpenEXR and RGBE formats. Through a series of experiments, we confirm that the proposed method decreases the volume of compressed data while maintaining the visual quality of the reconstructed LDR images.

We conducted a large-scale study of human perceptual quality judgments of High Dynamic Range (HDR) and Standard Dynamic Range (SDR) videos subjected to scaling and compression levels and viewed on three different display devices. While conventional expectations are that HDR quality is better than SDR quality, we have found subject preference of HDR versus SDR depends heavily on the display device, as well as on resolution scaling and bitrate. To study this question, we collected more than 23,000 quality ratings from 67 volunteers who watched 356 videos on OLED, QLED, and LCD televisions, and among many other findings, observed that HDR videos were often rated as lower quality than SDR videos at lower bitrates, particularly when viewed on LCD and QLED displays. Since it is of interest to be able to measure the quality of videos under these scenarios, e.g. to inform decisions regarding scaling, compression, and SDR vs HDR, we tested several well-known full-reference and no-reference video quality models on the new database. Towards advancing progress on this problem, we also developed a novel no-reference model called HDRPatchMAX, that uses a contrast-based analysis of classical and bit-depth features to predict quality more accurately than existing metrics.

Perceptual brightness-based inverse tone mapping for high dynamic range imaging