Systems and methods for generating synthetic training data
By generating a synthetic training data system that automatically copies and converts 3D point cloud data, the problem of time-consuming manual annotation in machine learning algorithm training is solved, the quantity and quality of training data are improved, and the training efficiency of object detection algorithms is enhanced.
CN113830101BActive Publication Date: 2026-07-17DENSO CORP
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- DENSO CORP
- Filing Date
- 2021-06-22
- Publication Date
- 2026-07-17
Smart Images
Figure CN113830101B_ABST
Abstract
The present disclosure relates to systems and methods for generating synthetic training data. A system for generating synthetic training data (185) can include one or more processors (110), and a memory (170) in communication with the one or more processors and having a receiving module (171), a duplicating module (172), and an inserting module (173). The modules have instructions that, when executed by the one or more processors, cause the one or more processors to receive original training data (182) that is in the form of a three-dimensional point cloud and has one or more original objects (183) formed by at least a portion of the three-dimensional point cloud, the one or more original objects annotated with original annotation data (506A, 506B), duplicate one of the one or more original objects to generate a synthetic object (187), and insert the synthetic object within the original training data to generate the synthetic training data.
Need to check novelty before this filing date? Find Prior Art