Back to Search
Document details
AHG8: On combined pre- and post-processing and ROI-based adaptive QP for machine vision
Abstract
This contribution reports the performance of enabling both pre- and post-processing and ROI-based adaptive QP for machine vision. According to experimental results, it is reported that when both pre- and post-processing and ROI-based adaptive QP in AGH8 software are enabled, the combined method achieves % (RA), % (LD) and % (AI) BD-rate savings for object detection task on the SFU-HW dataset under VCM CTC, which is higher than individually enabling pre- and post-processing or ROI-based adaptive QP for machine vision.
Cited:
ARCHIVE
:
._.DS_Store
._JVET-AI0196.pptx
._JVET-AI0196_JOINT_vtm_230_test1_JVET-AH2031-v1_anchor_time_vtm233_prepost_roiaqp.xlsm
._JVET-AI0196_JOINT_vtm_230_test1_JVET-AH2031-v1_anchor_time_vtm233_prepost_roiaqp_sfuld_time.xlsm
._JVET-AI0196_PREPOST_vtm_230_test3_JVET-AH2031-v1_anchor_time_vtm233_prepost.xlsm
._JVET-AI0196_ROIAQP_vtm_230_test2_JVET-AH2031-v1_anchor_time_vtm233_roiaqp.xlsm
._JVET-AI0196_TR.docx
._JVET-AI0196_v5.docx
._~$JVET-AI0196.pptx
._~$JVET-AI0196_JOINT_vtm_230_test1_JVET-AH2031-v1_anchor_time_vtm233_prepost_roiaqp.xlsm
._~$JVET-AI0196_PREPOST_vtm_230_test3_JVET-AH2031-v1_anchor_time_vtm233_prepost.xlsm
._~$JVET-AI0196_ROIAQP_vtm_230_test2_JVET-AH2031-v1_anchor_time_vtm233_roiaqp.xlsm
.DS_Store
JVET-AI0196.pptx
JVET-AI0196_JOINT_vtm_230_test1_JVET-AH2031-v1_anchor_time_vtm233_prepost_roiaqp.xlsm
JVET-AI0196_JOINT_vtm_230_test1_JVET-AH2031-v1_anchor_time_vtm233_prepost_roiaqp_sfuld_time.xlsm
JVET-AI0196_PREPOST_vtm_230_test3_JVET-AH2031-v1_anchor_time_vtm233_prepost.xlsm
JVET-AI0196_ROIAQP_vtm_230_test2_JVET-AH2031-v1_anchor_time_vtm233_roiaqp.xlsm
JVET-AI0196_TR.docx
JVET-AI0196_v5.docx
~$JVET-AI0196.pptx
~$JVET-AI0196_JOINT_vtm_230_test1_JVET-AH2031-v1_anchor_time_vtm233_prepost_roiaqp.xlsm
~$JVET-AI0196_PREPOST_vtm_230_test3_JVET-AH2031-v1_anchor_time_vtm233_prepost.xlsm
~$JVET-AI0196_ROIAQP_vtm_230_test2_JVET-AH2031-v1_anchor_time_vtm233_roiaqp.xlsm
Citation