US 12,339,950 B2
Weak password detection method and device based on deep learning, and electronic device
Fan Mo, Hangzhou (CN); Chenyue Yu, Hangzhou (CN); Yuan Fan, Hangzhou (CN); and Bo Liu, Hangzhou (CN)
Assigned to DBAPPSECURITY CO., LTD, Hangzhou (CN)
Appl. No. 18/022,751
Filed by DBAPPSECURITY CO., LTD, Hangzhou (CN)
PCT Filed Apr. 28, 2021, PCT No. PCT/CN2021/090394
§ 371(c)(1), (2) Date Feb. 23, 2023,
PCT Pub. No. WO2022/041815, PCT Pub. Date Mar. 3, 2022.
Claims priority of application No. 202010862802.9 (CN), filed on Aug. 25, 2020.
Prior Publication US 2023/0315835 A1, Oct. 5, 2023
Int. Cl. G06F 21/46 (2013.01)
CPC G06F 21/46 (2013.01) 13 Claims
OG exemplary drawing
 
1. A weak password detection method based on deep learning, comprising:
acquiring a password character string to be detected;
processing, by applying a fully trained weak password detection model, the password character string to be detected to obtain a strong/weak password classification label of the password character string to be detected, wherein the fully trained weak password detection model is obtained by training a deep learning model with a password character string as an input and a strong/weak password classification label corresponding to the password character string as a supervision; and
marking the password character string to be detected as a weak password under a condition that the strong/weak password classification label of the password character string to be detected is a weak password label;
after acquiring the password character string to be detected, the method further comprises:
pre-processing the password character string to be detected, to obtain a password character string to be detected after pre-processing;
the pre-processing the password character string to be detected, to obtain the password character string to be detected after pre-processing further comprises:
performing one-hot encoding for the password character string to be detected, and converting each character in the password character string to be detected into a code with a length denoted as N;
combining the password character string to be detected into a matrix of M rows and N columns from top to bottom according to an order of each character in the password string to be detected, wherein M is the number of characters in the password character string to be detected;
normalizing the matrix of M rows and N columns to obtain a matrix of X rows and N columns, wherein X is a preset number of rows;
performing an image scaling process on the matrix of X rows and N columns to obtain a matrix of X rows and N columns with a preset image size; and
taking the matrix of X rows and N columns with the preset image size as the password character string to be detected after pre-processing.