JVET-AC0079 [AHG15] Effect of the perceptual QP adaptation (QPA) on machine task performance [C. Kim, D. Gwak, J. Lim (LGE)]

This contribution reports the impact of the perceptual QP adaptation (QPA) on machine task performance. VTM 12.0 with QPA enabled configuration is used for encoding open image dataset that was used in the evaluation of the VCM CfP and object detection was performed using the decoded results. Experimental results reportedly show 15.17% BD rate gain for object detection.

Mainly effective for high QPs. It was reported by another participant that similar experiments were performed with other data sets, where in some cases gains, but sometimes also losses occurred.

It was suggested to study how this would combine with other approaches of VCM-related encoder optimization. In the current contribution, the VTM anchor was improved.

Further study was recommended.

It was suggested to have a BoG (coordinated by C. Hollmann and S. Liu) to meet after the joint meeting with WG 4 which should further discuss testing condition and a plan for software development.

JVET-AD0138 [AHG8] QPA with low activity threshold for machine task [C. Kim, D. Gwak, J. Lim (LGE)]

This contribution proposes a low activity threshold for QPA to reduce bit rate increases in less important areas for machine task. From experimental results, 1.19% performance improvement is observed on average.

At the previous JVET meeting, the impact of the perceptual QP adaptation (QPA) on object detection task performance for Open Images dataset was reported JVET-AC0079. This contribution reports on the object detection performance of the perceptual QP adaptation (QPA) on the SFU-HW video dataset, as well as the object detection performance using QPA with a low activity threshold.

Average performance improvement of 1.19% is reported compared to QPA which is in the software repository.

It is however commented that the results are inhomogeneous: In some cases, the results are better compared to current QPA, in other cases worse. Fluctuations are significant.

Further study was recommended to make the improvement more homogeneous.

It was also commented that BD rate changes in the range of 1% are not indicating much in this activity.

JVET-AC0351 BoG report on non-normative optimization for machine [C. Hollmann, S. Liu (BoG coordinators)]

This document contains the report of the BoG on non-normative optimization for machine. The BoG met on Wednesday 18 January 2023 at 1520-1720 UTC, discussing subjects related to non-normative optimization for machine consumption of coded video content, including:

• Common test conditions
  • Test sequences
  • RA/LD/AI configurations
  • QP points and BD-rate calculation
• Cross-check procedure
• Reference software
  • Where to setup the repository
  • What to be included in the initial version reference software
• Other aspects

The BoG recommended to:

• Align common test conditions with WG 4 VCM (including test sequences, test configurations, QP points, etc.)
• Further discuss on QP point selection for calculation of the BD-rate used for tool evaluation.
• Align crosscheck procedure with WG 4 VCM.
• Further discuss on crosscheck procedure for neural network-based encoding as well as processing tools.
• Setup repository for software development and test.
• Integrate technologies proposed in JVET-AB0275 and JVET-AC0086 to the reference software (VTM12.0) and share them in git.

Common test conditions

Document discussed: JVET-AC0073 AHG15: On common test conditions for optimization of encoders and receiving systems for machine analysis of coded video content [S. Liu (Tencent), C. Hollmann (Ericsson)]

This document describes common test conditions (CTC), reference software and configurations, reporting and other information to be used for experimenting and evaluating technologies for encoders and receiving systems for...