The purpose of this document is to provide descriptions of three core experiments on MPEG Compact Descriptors for Video Analysis (CDVA) (see N15339, “Call for Proposals for Compact Descriptors for Video Analysis (CDVA) - Search and Retrieval”, June 2015, Warsaw (PL)) based on the proposals submitted at 114th MPEG meeting and the submission to the 115th meeting. In addition to the CEs defined in the earlier version of this document (N16065, Description of Core Experiments on CDVA, February 2016, San Diego, US), the submissions to the 115th meeting showed that predictive coding of descriptors has the potential to improve performance while limiting descriptor size. However, the details of applying predictive coding still need to be investigated.
The results of experiments are scheduled to be discussed during the CDVA AHG meetings before 116th MPEG meeting.
The report of each CE should include; (1) a comparison between the tested solutions, (2) a recommendation to the AhG based on the results of the CE.
The results according to the evaluation criteria described below shall be reported (in particular those defined by the evaluation framework - N15729, Evaluation Framework for Compact Descriptors for Video Analysis - Search and Retrieval – Version 2.0, October 2015, Geneva, CH]), but results based on other evaluation criteria may be included.
