US 11,966,822 B2
Feature processing for machine learning
Chun Lei Xu, Xi'an (CN); Si Er Han, Xi'an (CN); Shi Bin Liu, Xi'an (CN); Yi Shao, Xi'an (CN); Lei Tian, Xi'an (CN); Hao Zheng, Xi'an (CN); and Jia Rui Wang, Xi'an (CN)
Assigned to International Business Machines Corporation, Armonk, NY (US)
Filed by International Business Machines Corporation, Armonk, NY (US)
Filed on Sep. 29, 2020, as Appl. No. 17/035,699.
Prior Publication US 2022/0101183 A1, Mar. 31, 2022
Int. Cl. G06F 18/22 (2023.01); G06F 18/213 (2023.01); G06N 20/00 (2019.01); G06V 10/77 (2022.01)
CPC G06N 20/00 (2019.01) [G06F 18/213 (2023.01); G06F 18/22 (2023.01); G06V 10/7715 (2022.01)] 18 Claims
OG exemplary drawing
 
1. A computer-implemented method for feature processing, comprising:
projecting, by one or more processing units, two input image recognition features selected from multiple features of each sample in a sample set to one resulting feature based on a specified curve, wherein the projecting the two input image recognition features comprises:
projecting, by one or more processing units, a two-dimensional data point to a one-dimensional data point on the specified curve, wherein two values of two axes of the two-dimensional data point in a coordinate system are respectively determined from the two input image recognition features, and the one-dimensional data point is a data point on the specified curve which has the shortest distance to the two-dimensional data point as an orthogonal distance, and
determining, by one or more processing units, an arc length from the one-dimensional data point to a start point in the specified curve as the resulting feature;
updating, by one or more processing units, the sample set by replacing the two input image recognition features with the one resulting feature for each sample in the sample set; and
repeating, by one or more processing units, the projecting and the updating for the sample set until the number of features of each sample in the sample set reaches a predetermined criterion.