Patents by Inventor Zhifeng HAO
Zhifeng HAO has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).
-
Publication number: 20240104006Abstract: Disclosed is a method for automatically generating interactive test cases. The method comprises: after a UI of an application program is displayed, traversing all views in a view tree corresponding to the UI of the application program, and recording a path, in the view tree, of each view therein that can be clicked on, so as to obtain a set of path information, in the view tree, of all the views that can be clicked on in the UI; and respectively generating a corresponding test case for each piece of path information in the set: in the test case, according to path information, in the view tree, of a view under test, finding the view in the UI interface of the application program, and triggering a click event therefore, that is, completing a click interaction test on the view. In the present invention, there is no strict requirements for the type and the running environment of an application program.Type: ApplicationFiled: October 8, 2021Publication date: March 28, 2024Applicant: SOUTH CHINA UNIVERSITY OF TECHNOLOGYInventors: Han HUANG, Jie CAO, Lei YE, Fangqing LIU, Zhifeng HAO
-
Patent number: 11941449Abstract: The field of high-speed data acquisition and network data processing, and particularly relates to an Ethernet data stream recording method, an Ethernet data stream recording system, and an Ethernet data stream recording device for a high-speed data acquisition system. It is intended to solve problems such as a low utilization rate of CPU, poor system compatibility, difficulty in packaging and deployment and low reliability of system transmission of the traditional high-speed data acquisition system.Type: GrantFiled: May 25, 2020Date of Patent: March 26, 2024Assignees: Institute of Automation, Chinese Academy of Sciences, Guangdong Institute of Artificial Intelligence and Advanced ComputingInventors: Zhifeng Lv, Jie Hao, Jun Liang, Lin Shu, Meiting Zhao, Yafang Song, Qiuxiang Fan
-
Patent number: 11720477Abstract: The present invention provides a unit testing method based on automatic generation of path coverage test cases. First, obtain a control flow graph of a program to be tested is obtained. Then, an executable code is executed in the generated control flow graph based on of an automatically generated test case, and meanwhile, a fitness value is calculated and acquired based on of an execution result of the executable code, and a sub-node is selected to continue repeating the above process, until a terminal node in the control flow graph is found, and finally a path marker is generated and the fitness value corresponding to the path is obtained. Then, an automatic test case generation algorithm is executed, and the algorithm constantly automatically generates test cases based on of the returned fitness value, and exits when the path is completely covered, or a set execution is timed out.Type: GrantFiled: October 31, 2018Date of Patent: August 8, 2023Assignee: SOUTH CHINA UNIVERSITY OF TECHNOLOGYInventors: Han Huang, Muming Lian, Fangqing Liu, Zhongming Yang, Zhifeng Hao
-
Patent number: 11714678Abstract: An intelligent scheduling method for supporting process task quantity splitting, which may relax the limit on the number of parallel machines for overdue task lists under the constraint of using as few parallel machines as possible, and split time-consuming process task quantities according to the operating status of machines in different periods.Type: GrantFiled: October 24, 2018Date of Patent: August 1, 2023Assignee: SOUTH CHINA UNIVERSITY OF TECHNOLOGYInventors: Han Huang, Junpeng Su, Xueqiang Li, Zhifeng Hao
-
Publication number: 20230185703Abstract: Disclosed is an automatic parsing and path analysis method for a unit test code structure. The method comprises: acquiring compiled byte codes according to a language of a test program; traversing the complied byte codes, making instrumentation codes respectively in front of important statements, and acquiring node information and a small-segment path set (SSPS); analyzing the SSPS, replacing a part therein comprising nesting to obtain a SSPS excluding nesting as a basis, initializing a path table among the nodes, updating the path table by utilizing a depth-first algorithm, and obtaining path sets according to the path table; if all the path sets have been updated, returning to continuously update the path table; and outputting the acquired path sets and a program flowchart CFG obtained by analysis. The method above is capable of acquiring the path sets efficiently, thereby improving the capability of processing path analysis in actual software unit test.Type: ApplicationFiled: October 28, 2021Publication date: June 15, 2023Applicant: SOUTH CHINA UNIVERSITY OF TECHNOLOGYInventors: Fangqing LIU, Han HUANG, Xiao LING, Feng LIN, Jie CAO, Shaoyang ZHUANG, Zhifeng HAO
-
Publication number: 20230157845Abstract: A femoral neck measuring method includes: measuring an affected limb after a femoral stem prosthesis is installed through an optical localization probe, and calculating an actual length of the femoral neck needed by the affected limb according to an osteotomy surface of the affected limb determined after the femoral stem prosthesis is installed; determining a femoral ball prosthesis and a femoral neck prosthesis to be selected according to the calculated actual length of the femoral neck and a length of a femoral neck of a healthy limb measured before surgery; and measuring a length of the affected limb after the femoral ball prosthesis and the femoral neck prosthesis are implanted, and restraining a length difference between the healthy limb and the affected limb by adjusting a length of the femoral neck of the affected limb.Type: ApplicationFiled: October 26, 2022Publication date: May 25, 2023Applicants: Shantou University, The First Affiliated Hospital of Shantou University Medical CollegeInventors: Zhun Fan, Weibo Ning, Jiazi Liu, Jiaqi Zhu, Guijie Zhu, Jun Hu, Hongjiang Chen, Jiacheng Liu, Zhifeng Hao
-
Patent number: 11604792Abstract: The invention discloses a method and a device for constructing a SQL statement based on reinforcement learning, wherein the method includes: initializing an actor-critic network parameter; acquiring a sequence pair of natural language and real SQL statement from a data set; inputting a natural language sequence into an actor network encoder, and inputting a real SQL sequence into a critic network encoder; using an encoded hidden state as an initialized hidden state of a corresponding decoder; gradually predicting, by an actor network decoder, a SQL statement action, and inputting the SQL statement action to a critic network decoder and an environment to obtain a corresponding reward; and using a gradient descent algorithm to update the network parameters, and obtaining a constructing model of the natural language to the SQL statement after repeated iteration training.Type: GrantFiled: March 20, 2020Date of Patent: March 14, 2023Assignee: GUANGDONG UNIVERSITY OF TECHNOLOGYInventors: Ruichu Cai, Boyan Xu, Zhihao Liang, Zijian Li, Zhifeng Hao, Wen Wen, Bingfeng Chen
-
Patent number: 11507049Abstract: The present invention discloses a method for detecting abnormity in an unsupervised industrial system based on deep transfer learning. Labeled machine sensor sequence data from a source domain and unlabeled sensor sequence data from a target domain are used in the present invention to train an industrial system abnormal detection model with good generalization ability, and the industrial system abnormal detection model is trained and tested to finally generate a trained industrial system abnormity discrimination model. Using the model, received machine sensor sequence data can be analyzed and whether a machine is abnormal is discriminated.Type: GrantFiled: November 12, 2019Date of Patent: November 22, 2022Assignee: GUANGDONG UNIVERSITY OF TECHNOLOGYInventors: Ruichu Cai, Zijian Li, Wen Wen, Zhifeng Hao, Lijuan Wang, Bingfeng Chen, Boyan Xu, Junfeng Li
-
Patent number: 11417147Abstract: The present invention provides an angle interference resistant and occlusion interference resistant fast face recognition method, comprising: first collecting a training set of images which have been detected and cropped, adjusting and expanding the training set, and conducting standardized pre-processing; inputting the same into a constructed neural network for training, and saving a parametric model; adjusting test data to a suitable size and number of channels, and also conducting standardized pre-processing; inputting the same into a prediction network to obtain feature vectors of face images; and determining whether two faces are from the same person by calculating a distance between the feature vectors of the face images.Type: GrantFiled: December 31, 2018Date of Patent: August 16, 2022Assignee: SOUTH CHINA UNIVERSITY OF TECHNOLOGYInventors: Han Huang, Di Liu, Zhifeng Hao
-
Patent number: 11263434Abstract: Disclosed is a fast side-face interference resistant face detection method, in which a user selects an ordinary image, uses a deep neural network to extract image features, and then determines an exact location of a face. A training method for face detection uses a pure data-driven manner, uses an ordinary face image and a face boundary box as inputs, uses mirror symmetry and Gaussian filtering to perform data augmentation, and uses migration learning and hard example mining to enhance training effects. After a face image is read, the image is firstly scaled, and then placed into the deep neural network to extract features, and generate a plurality of face likelihood boxes and confidence scores of the face likelihood boxes, and finally the most appropriate face likelihood box is selected in a non-maximum suppression manner. No specific requirements are set on an angle of the face image, and a detection effect of a side face is still very obvious.Type: GrantFiled: November 15, 2018Date of Patent: March 1, 2022Assignee: SOUTH CHINA UNIVERSITY OF TECHNOLOGYInventors: Han Huang, Zilong Li, Zhifeng Hao
-
Patent number: 11074052Abstract: The invention provides an automatic compiling method for graphical programming, which mainly comprises the following steps of allocating a graphical program memory, allocating a graphical program thread, analyzing a graphical program storage structure and generating a graphical program executable file. An executable file corresponding to a graphical program is generated based on automatic compiling of graphical programming, a user freely combines graphical modules according to functional requirements to form a program, an automatic compiling method for graphical programming is used for compiling the graphical program, and a file which can be directly operated in a controller is generated.Type: GrantFiled: November 16, 2017Date of Patent: July 27, 2021Assignee: South China University of TechnologyInventors: Han Huang, Liang Qin, Zhanning Liang, Zhifeng Hao, Zhun Fan
-
Publication number: 20210209008Abstract: The present invention provides a unit testing method based on automatic generation of path coverage test cases. First, obtain a control flow graph of a program to be tested is obtained. Then, an executable code is executed in the generated control flow graph based on of an automatically generated test case, and meanwhile, a fitness value is calculated and acquired based on of an execution result of the executable code, and a sub-node is selected to continue repeating the above process, until a terminal node in the control flow graph is found, and finally a path marker is generated and the fitness value corresponding to the path is obtained. Then, an automatic test case generation algorithm is executed, and the algorithm constantly automatically generates test cases based on of the returned fitness value, and exits when the path is completely covered, or a set execution is timed out.Type: ApplicationFiled: October 31, 2018Publication date: July 8, 2021Applicant: SOUTH CHINA UNIVERSITY OF TECHNOLOGYInventors: Han HUANG, Muming LIAN, Fangqing LIU, Zhongming YANG, Zhifeng HAO
-
Patent number: 11037340Abstract: Disclosed is an automatic obstacle avoidance optimization method for a connecting line of graphical programming software. A breadth-first search method is adopted to search an optimal connecting line path, and a result is displayed in a front-end interface. In a generated connecting line, a user can drag the connecting line with a mouse to regulate a position of the connecting line. At the time, the mouse is viewed as an unavoidable point, that is, the connecting line starts from a starting point of the connecting line to the unavoidable point and then reaches an end point of the connecting line. By means of the method, high instantaneity requirements of the user can be met, and when the user drags the connecting line with the mouse, a new connecting line is generated in real time without a stuck phenomenon.Type: GrantFiled: November 15, 2017Date of Patent: June 15, 2021Assignee: South China University of TechnologyInventors: Han Huang, Hongyue Wu, Zhifeng Hao, Xiaowei Yamg
-
Patent number: 10963224Abstract: The present invention discloses a graphical programming control and storage system, which includes a central control module, a display module connected to the central control module, a message response module, a graphical programming inter-component operating module, a graphical programming intra-component operating module, an item attribute and control module, a user-defined component module, an item persistence module, and a compilation module; and the central control module is responsible for processing, storing, and returning a delivered information and result. The system enables a user to complete complex programming by dragging the graphical programming components, and finally generates a formulated language or executable program, thus implementing graphical programming. The present invention can implement graphical programming control and storage by the solution above, enabling visible and intuitive programming, and improving the working efficiency of programmers.Type: GrantFiled: October 26, 2017Date of Patent: March 30, 2021Assignee: SOUTH CHINA UNIVERSITY OF TECHNOLOGYInventors: Han Huang, Hu Wang, Yihui Liang, Yichen Sheng, Zhifeng Hao
-
Publication number: 20210049041Abstract: The invention discloses an intelligent scheduling method for supporting process task quantity splitting, which may relax the limit on the number of parallel machines for overdue task lists under the constraint of using as few parallel machines as possible, and split time-consuming process task quantity according to the operating status of machines in different periods. Compared with the current commonly used production planning and scheduling methods, which generally only allocate the task quantity of a single process to a single machine, and may not achieve the splitting of the task quantity, and when the production task quantity is large and the number of idle machines of the same type is large, it will cause waste of machine resources and low production efficiency, the present invention may support the asynchronous starting mechanism of the same type of machines with asynchronous operating status.Type: ApplicationFiled: October 24, 2018Publication date: February 18, 2021Inventors: Han HUANG, Junpeng SU, Xueqiang LI, Zhifeng HAO
-
Publication number: 20200410214Abstract: The present invention provides an angle interference resistant and occlusion interference resistant fast face recognition method, comprising: first collecting a training set of images which have been detected and cropped, adjusting and expanding the training set, and conducting standardized pre-processing; inputting the same into a constructed neural network for training, and saving a parametric model; adjusting test data to a suitable size and number of channels, and also conducting standardized pre-processing; inputting the same into a prediction network to obtain feature vectors of face images; and determining whether two faces are from the same person by calculating a distance between the feature vectors of the face images.Type: ApplicationFiled: December 31, 2018Publication date: December 31, 2020Applicant: SOUTH CHINA UNIVERSITY OF TECHNOLOGYInventors: Han HUANG, Di LIU, Zhifeng HAO
-
Publication number: 20200410212Abstract: Disclosed is a fast side-face interference resistant face detection method, in which a user selects an ordinary image, uses a deep neural network to extract image features, and then determines an exact location of a face. A training method for face detection uses a pure data-driven manner, uses an ordinary face image and a face boundary box as inputs, uses mirror symmetry and Gaussian filtering to perform data augmentation, and uses migration learning and hard example mining to enhance training effects. After a face image is read, the image is firstly scaled, and then placed into the deep neural network to extract features, and generate a plurality of face likelihood boxes and confidence scores of the face likelihood boxes, and finally the most appropriate face likelihood box is selected in a non-maximum suppression manner. No specific requirements are set on an angle of the face image, and a detection effect of a side face is still very obvious.Type: ApplicationFiled: November 15, 2018Publication date: December 31, 2020Applicant: SOUTH CHINA UNIVERSITY OF TECHNOLOGYInventors: Han HUANG, Zilong LI, Zhifeng HAO
-
Publication number: 20200301924Abstract: The invention discloses a method and a device for constructing a SQL statement based on reinforcement learning, wherein the method includes: initializing an actor-critic network parameter; acquiring a sequence pair of natural language and real SQL statement from a data set; inputting a natural language sequence into an actor network encoder, and inputting a real SQL sequence into a critic network encoder; using an encoded hidden state as an initialized hidden state of a corresponding decoder; gradually predicting, by an actor network decoder, a SQL statement action, and inputting the SQL statement action to a critic network decoder and an environment to obtain a corresponding reward; and using a gradient descent algorithm to update the network parameters, and obtaining a constructing model of the natural language to the SQL statement after repeated iteration training.Type: ApplicationFiled: March 20, 2020Publication date: September 24, 2020Applicant: GUANGDONG UNIVERSITY OF TECHNOLOGYInventors: Ruichu CAI, Boyan XU, Zhihao LIANG, Zijian LI, Zhifeng HAO, Wen WEN, Bingfeng CHEN
-
Publication number: 20200286268Abstract: Disclosed is an automatic obstacle avoidance optimization method for a connecting line of graphical programming software. A breadth-first search method is adopted to search an optimal connecting line path, and a result is displayed in a front-end interface. In a generated connecting line, a user can drag the connecting line with a mouse to regulate a position of the connecting line. At the time, the mouse is viewed as an unavoidable point, that is, the connecting line starts from a starting point of the connecting line to the unavoidable point and then reaches an end point of the connecting line. By means of the method, high instantaneity requirements of the user can be met, and when the user drags the connecting line with the mouse, a new connecting line is generated in real time without a stuck phenomenon.Type: ApplicationFiled: November 15, 2017Publication date: September 10, 2020Applicant: South China University of TechnologyInventors: Han HUANG, Hongyue WU, Zhifeng HAO, Xiaowei YAMG
-
Publication number: 20200150622Abstract: The present invention discloses a method for detecting abnormity in an unsupervised industrial system based on deep transfer learning. Labeled machine sensor sequence data from a source domain and unlabeled sensor sequence data from a target domain are used in the present invention to train an industrial system abnormal detection model with good generalization ability, and the industrial system abnormal detection model is trained and tested to finally generate a trained industrial system abnormity discrimination model. Using the model, received machine sensor sequence data can be analyzed and whether a machine is abnormal is discriminated.Type: ApplicationFiled: November 12, 2019Publication date: May 14, 2020Applicant: GUANGDONG UNIVERSITY OF TECHNOLOGYInventors: Ruichu CAI, Zijian LI, Wen WEN, Zhifeng HAO, Lijuan WANG, Bingfeng CHEN, Boyan XU, Junfeng LI