US 12,307,612 B2
Apparatus and method for creating intelligent special effects based on object recognition
Oh Young Song, Seoul (KR); Soo Mi Choi, Seoul (KR); Bo Seong Kim, Seoul (KR); and Jin Seong Yang, Seoul (KR)
Assigned to INDUSTRY ACADEMY COOPERATION FOUNDATION OF SEJONG UNIVERSITY, Seoul (KR)
Filed by INDUSTRY ACADEMY COOPERATION FOUNDATION OF SEJONG UNIVERSITY, Seoul (KR)
Filed on Feb. 28, 2023, as Appl. No. 18/115,161.
Claims priority of application No. 10-2022-0165636 (KR), filed on Dec. 1, 2022.
Prior Publication US 2024/0185537 A1, Jun. 6, 2024
Int. Cl. G06T 19/00 (2011.01); G06T 19/20 (2011.01); G06V 10/26 (2022.01); G06V 10/774 (2022.01); G06V 20/20 (2022.01); G06V 20/64 (2022.01)
CPC G06T 19/006 (2013.01) [G06T 19/20 (2013.01); G06V 10/26 (2022.01); G06V 10/774 (2022.01); G06V 20/20 (2022.01); G06V 20/647 (2022.01); G06T 2219/2016 (2013.01)] 8 Claims
OG exemplary drawing
 
1. An apparatus for creating intelligent special effects based on object recognition comprising:
a communication module for receiving a photographed video of an object;
a memory storing a program for creating special effect information from the received video; and
a processor for executing the program stored in the memory,
wherein the program inputs an object video acquired from a camera to an object recognition model and extracts an object image to which an attribute of each object is matched,
weight information of each object is derived by inputting size information of the object image for each attribute of each object to a weight estimation model, and
special effect control information mapped to each object is determined based on the weight information for each attribute of each object,
wherein in the weight estimation model, a multi-regression model is constructed for each attribute of each object so that the weight information of each object is output by taking actual horizontal and vertical length information of each object calculated by the program as an input.