Perceptually Optimized Bit-Allocation and Associated Distortion Measure for Block-Based Image or Video Coding

Christian R. Helmrich, Senior Member, IEEE, Sebastian Bosse, Mischa Siekmann, Heiko Schwarz, Detlev Marpe, Fellow, IEEE, and Thomas Wiegand, Fellow, IEEE

The above paper has been presented and published at the IEEE Data Compression Conference (DCC) in Snowbird, Utah in March 2019.  This web page contains supplementary information on the publication, particularly visual demonstrations.

  1. Corrections to the DCC submitted version of the paper as of March 2019

  2. Visual demonstrations of the perceptual bit-allocation in the VVC encoder

  1. Corrections to the Data Compression Conference submitted version as of March 2019

    Aside from minor issues with the text, particularly in Section 5.2, the parameter Omax in Section 4.2 was not introduced, and a reference to the HEVC standard text [1] was missing in the submitted manuscript.  A first corrected version of the paper is available here: link to document DCC manuscript

    For the published version, the following additional corrections have been made:

    • Sec. 4.1: mentioned empirical origin of amin

    • Sec. 5: fixed definition of d (27+BD instead of 217)

    • Sec. 5.2: mentioned random-access configuration (GOP size 16)

    • Sec. 6: mentioned this demo page.

    As of March 31, 2019, the only known issue is a typographical error in equation (12): it should read Δ2k instead of Δk.


  2. Subjective demonstrations of the perceptual bit-allocation in VVC still-image coding

    The following illustrations serve as a demonstration of the visual benefit of using the perceptually optimized quantization parameter adaptation (QPA) in a transform-based still-image codec like HEVC [1].  The basic coding algorithm used for this demonstration is draft 3 of the Versatile Video Coding (VVC) specification [2], as implemented by the VTM3.0 reference software [3] into which our QPA method has been integrated. Since only single images are utilized for this demonstration, the VVC codec was configured to apply only “still-image” Intra-picture prediction.

    The presented images were transcoded, with visual transparency, to high-bit-rate JPEG in order to limit the download durations for the viewers. Differences between the coded pictures are mostly visible in low-contrast regions, so viewing in low background-lighting conditions is advised.

    BQTerrace_original - Click to view full-sized image.
    BQTerrace, uncoded input (HD, 1920×1080, lossless size: 4989 KB)

    BQTerrace_qp32_QPAoff - Click to view full-sized image.
    BQTerrace, VTM 3.0.1 without QPA, base QP 32 (HD, 1920×1080, coded size: 96.7 KB)

    BQTerrace_qp29_QPAon - Click to view full-sized image.
    BQTerrace, VTM 3.0.1 with QPA, base QP 29 (HD, 1920×1080, coded size: 98.5 KB)

    BasketballDrive68_original - Click to view full-sized image.
    BasketballDrive, frame 68, uncoded input (HD, 1920×1080, lossless size: 5123 KB)

    BasketballDrive68_qp30_QPAoff - Click to view full-sized image.
    BasketballDrive, frame 68, VTM 3.0.1 without QPA, base QP 30 (HD, 1920×1080, coded size: 51.5 KB)

    BasketballDrive68_qp30_QPAon - Click to view full-sized image.
    BasketballDrive, frame 68, VTM 3.0.1 with QPA, base QP 30 (HD, 1920×1080, coded size: 49.2 KB)

    Kodim15_original - Click to view full-sized image.
    Kodak Image 15, uncoded input (768×512, lossless size: 755 KB)

    Kodim15_qp28_QPAoff - Click to view full-sized image.
    Kodak Image 15, VTM 3.0.1 without QPA, base QP 28 (768×512, coded size: 22.2 KB)

    Kodim15_qp29_QPAon - Click to view full-sized image.
    Kodak Image 15, VTM 3.0.1 with QPA, base QP 29 (768×512, coded size: 22.4 KB)

    ParkScene_original - Click to view full-sized image.
    ParkScene, uncoded input (HD, 1920×1080, lossless size: 4911 KB)

    ParkScene_qp29_QPAoff - Click to view full-sized image.
    ParkScene, VTM 3.0.1 without QPA, base QP 29 (HD, 1920×1080, coded size: 95.9 KB)

    ParkScene_qp30_QPAon - Click to view full-sized image.
    ParkScene, VTM 3.0.1 with QPA, base QP 30 (HD, 1920×1080, coded size: 92.9 KB)

Please note that the layout or content of this web page may change.  If it does, the date in the following line will be updated.

Christian R. Helmrich, March 31, 2019



References

[1]  ITU-T, Recommendation H.265 and ISO/IEC, Int. Standard 23008-2, “High efficiency video coding,” Geneva, Switzerland, Feb. 2018. Online:
       link to external web page http://www.itu.int/rec/T-REC-H.265

[2]  B. Bross, J. Chen, S. Liu, “Versatile Video Coding (Draft 3),” MPEG/JVET output document m45225/L1001, Macao, China, Dec. 2018. Online:
       link to external web page http://phenix.it-sudparis.eu/jvet/doc_end_user/current_document.php?id=4834

[3]  JVET, “VVCSoftware_VTM: VVC VTM reference software,” Dec. 2018. Online: link to external web page https://vcgit.hhi.fraunhofer.de/jvet/VVCSoftware_VTM/tags