Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
IMAGE ENCODING AND DECODING USING DEEP LEARNING-BASED INTER PREDICTION
Document Type and Number:
WIPO Patent Application WO/2022/031115
Kind Code:
A1
Abstract:
The present embodiment provides an image encoding/decoding method, which performs conventional inter prediction on the current block to generate a motion vector and prediction samples, and uses a deep learning-based video prediction network (VPN) on the basis of the motion vector, reference samples, the prediction samples, and the like to generate improved prediction samples for the current block, and thus improves encoding efficiency.

Inventors:
KANG JE WON (KR)
PARK SEUNG WOOK (KR)
LIM WHA PYEONG (KR)
Application Number:
PCT/KR2021/010425
Publication Date:
February 10, 2022
Filing Date:
August 06, 2021
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
HYUNDAI MOTOR CO LTD (KR)
KIA CORP (KR)
UNIV EWHA IND COLLABORATION (KR)
International Classes:
G06T9/00; G06N3/08; G06T3/40; G06T5/00; H04N19/513
Foreign References:
KR20190096281A2019-08-19
KR20190033403A2019-03-29
US20190306526A12019-10-03
US20180249158A12018-08-30
KR20200050433A2020-05-11
Attorney, Agent or Firm:
LEE, Chulhee (KR)
Download PDF:
View/Download PDF      PDF Help



 
Previous Patent: METHOD FOR OPERATING IAB NODE IN WIRELESS COMMUNICATION SYSTEM, AND DEVICE USING METHOD

Next Patent: POSITIVE ELECTRODE ACTIVE MATERIAL PRECURSOR AND METHOD FOR PREPARING SAME