US 11,989,927 B2
Apparatus and method for detecting keypoint based on deep learning using information change across receptive fields
Yong-Ju Cho, Daejeon (KR); Jeong-Il Seo, Daejeon (KR); Rehan Hafiz, Lahore (PK); Mohsen Ali, Lahore (PK); Muhammad Faisal, Lahore (PK); Usama Sadiq, Lahore (PK); and Tabasher Arif, Lahore (PK)
Assigned to ELECTRONICS AND TELECOMMUNICATIONS RESEARCH INSTITUTE, Daejeon (KR); and INFORMATION TECHNOLOGY UNIVERSITY (ITU), Lahore (PK)
Filed by ELECTRONICS AND TELECOMMUNICATIONS RESEARCH INSTITUTE, Daejeon (KR); and Information Technology University (ITU), Lahore (PK)
Filed on Dec. 30, 2021, as Appl. No. 17/565,901.
Claims priority of application No. 10-2021-0092530 (KR), filed on Jul. 14, 2021.
Prior Publication US 2023/0035307 A1, Feb. 2, 2023
Int. Cl. G06V 10/77 (2022.01); G06T 3/4007 (2024.01); G06V 10/44 (2022.01); G06V 10/82 (2022.01)
CPC G06V 10/7715 (2022.01) [G06T 3/4007 (2013.01); G06V 10/443 (2022.01); G06V 10/82 (2022.01)] 17 Claims
OG exemplary drawing
 
1. An apparatus for detecting a keypoint based on deep learning robust to scale changes based on information change across receptive fields, the apparatus comprising:
at least one processor; and
at least one memory including instructions that when executed by the at least one processor, implement:
a feature extractor that extracts a feature from an input image based on a pre-trained deep learning neural network;
an information accumulation pyramid part that outputs, from the feature, at least two filter responses corresponding to receptive fields having different scales;
an information change detector that calculates an information change between the at least two filter responses;
a keypoint detector that creates a score map having a keypoint probability of each pixel based on the information change; and
a continuous scale estimator that estimates a scale of a receptive field having a biggest information change for each pixel.