US 11,928,790 B2
Object recognition device, object recognition learning device, method, and program
Yukito Watanabe, Tokyo (JP); Jun Shimamura, Tokyo (JP); and Atsushi Sagata, Tokyo (JP)
Assigned to NIPPON TELEGRAPH AND TELEPHONE CORPORATION, Tokyo (JP)
Appl. No. 17/268,293
Filed by NIPPON TELEGRAPH AND TELEPHONE CORPORATION, Tokyo (JP)
PCT Filed Aug. 8, 2019, PCT No. PCT/JP2019/031469
§ 371(c)(1), (2) Date Feb. 12, 2021,
PCT Pub. No. WO2020/036124, PCT Pub. Date Feb. 20, 2020.
Claims priority of application No. 2018-152701 (JP), filed on Aug. 14, 2018.
Prior Publication US 2021/0303903 A1, Sep. 30, 2021
Int. Cl. G06T 3/40 (2006.01); G06F 16/583 (2019.01); G06N 3/045 (2023.01); G06N 3/084 (2023.01); G06N 20/00 (2019.01); G06V 10/764 (2022.01); G06V 10/82 (2022.01); G06V 30/24 (2022.01)
CPC G06T 3/4046 (2013.01) [G06F 16/583 (2019.01); G06N 3/045 (2023.01); G06N 3/084 (2013.01); G06N 20/00 (2019.01); G06V 10/764 (2022.01); G06V 10/82 (2022.01); G06V 30/2504 (2022.01)] 20 Claims
OG exemplary drawing
 
1. An object recognition learning device for recognizing an object captured on a query image serving as a recognition subject, the device comprising:
an acquirer configured to acquire, from the query image, an increased-resolution image, which is acquired by increasing the resolution of the query image, by performing pre-learned acquisition processing for increasing the resolution of an image;
a feature extractor configured to, using the increased-resolution image as input, extract a feature vector of the increased-resolution image by performing pre-learned extraction processing for extracting a feature vector of an image; and
a recognizer configured to recognize an object captured on the increased-resolution image on the basis of the feature vector of the increased-resolution image and outputs the recognized object as the object captured on the query image, wherein the acquisition processing of the acquirer is acquisition processing learned at least by associating an arbitrary image on which the object is captured with an image acquired by increasing the resolution of a low-resolution image corresponding to the arbitrary image, and during learning, the acquisition processing is optimized at the same time as the extraction processing.