AsianScientist (Apr. 19, 2018) – Scientists from China have used a virtual environment to train computers to understand objects in the real world. Their findings have been published in IEEE/CAA Journal of Automatica Sinica, a joint publication of the IEEE and the Chinese Association of Automation.
For computers to accurately recognize objects, they must process a huge amount of labeled data, in this case, images of objects with accurate annotations. A self-driving car, for instance, needs thousands of images of roads and cars to learn how to drive safely. Datasets therefore play a crucial role in the training and testing of computer vision systems.
“However, collecting and annotating images from the real world is too demanding in terms of labor and money investments,” wrote Wang Kunfeng, an associate professor at China’s State Key Laboratory for Management and Control for Complex Systems and the lead author on the paper.
To tackle this problem, Wang and his colleagues created a dataset called ParallelEye. ParallelEye was virtually generated by using commercially available computer software, primarily the video game engine Unity3D.
Using a map of Zhongguancun, one of the busiest urban areas in Beijing, China, as their reference, they recreated an urban setting by adding various buildings, cars and even different weather conditions. The researchers then placed a virtual camera on a virtual car, allowing it to drive around the virtual Zhongguancun and create datasets that are representative of the real world.
While their greatest performance increases came from incorporating ParallelEye datasets with real world datasets, the team has demonstrated that their method is capable of creating diverse sets of purely virtual images as well.
“Using the ParallelEye vision framework, massive and diversified images can be synthesized flexibly and this can help build more robust computer vision systems,” said Wang.
The research team’s proposed approach can be applied to many visual computing scenarios, including visual surveillance, medical image processing and biometrics. The team plans on creating an even larger set of virtual images, improving the realism of their virtual images and exploring the utility of virtual images for other computer vision tasks.
The article can be found at: Tian et al. (2018) Training and Testing Object Detectors with Virtual Images.
———
Source: Chinese Association of Automation; Photo: Shutterstock.
Disclaimer: This article does not necessarily reflect the views of AsianScientist or its staff.
