Unity Software Inc., a Video-game software developer, has developed synthetic image datasets that it claims can be used to create computer vision, artificial intelligence models faster and much lower cost.
Unity is primarily known for its 3D gaming engine that video game developers use to create more interactive gaming environments and virtual reality and augmented reality apps. However, its software relies on AI to render its virtual worlds. It can also create full-scale models of objects that can aid design in industries like architecture, engineering, and construction.
The new Unity Computer Vision Datasets are based on synthetic data. Synthetic data is the information that is generated artificially using algorithms. It is often used as a stand-in for test datasets of new products and tools, model validation, and AI model training.
According to Unity, its synthetic datasets are pretty helpful as the information within them is generated to meet many specific needs and conditions that are not available in existing or real data. Another significant benefit of synthetic data is that it eliminates the privacy pitfalls that arise from using images of real people and places.
However, a problem with synthetic data is that it must be created by an AI model first, which can be an expensive and time-consuming process. According to Unity, the Unity engine is used to create digital twins of objects of interest. Then they are placed in a variety of 3D environments with diverse lighting conditions, orientations, scale factors, and other parameters, called randomizers. These digital twins are then used to create high volumes of various training data for machine learning.
Synthetic data has the unlimited potential to produce datasets that are diverse in any dimension. It is perfectly labeled and is much less expensive to refresh. Unity is noticing that large-scale data diversity is more important for producing accurate AI models than limited sets of real data.
Computer Vision Datasets can support a wide variety of AI training use cases too. One of the most prominent examples is object detection, which can be applied to manufacturing, retail, and security scenarios. Unity believes that it can work with customers to ensure the datasets are produced in a way that meets the exact criteria for their needs. The company offers a tiered pricing model in which the price per image falls proportionally with the increased demand for more synthetic images.