"After achieving remarkable success in R&D, BasicAI has garnered global acclaim, including recognition from several Korean companies. At DataMatica, we are thrilled to join hands with exceptional companies like BasicAI."
Shih-Hung Chu (KR)
CEO, DataMatica
Established in 2020, DataMatica is a South Korean firm that provides state-of-the-art data processing services in the areas of autonomous driving, digital twin, HD map construction, and autonomous shuttle operation. DataMatica continues to grow by collaborating with companies, universities, organizations, local governments, and various departments, such as Inha University, Korea Smart Information Society Promotion Agency, Korea Transport Safety Authority, and Railway Technology Research Institute in Korea.
Challenges
Tracking team performances and costs in the team annotation can be difficult
In the process of team collaboration for data annotation, it is vital to accurately quantify team performance and costs. "At DataMatica, we faced a significant challenge in accurately determining costs before partnering with BasicAI platform", said CEO Shih-Hung Chu. Even with a designated budget plan outlining the cost of each object, variations in the number, type, and complexity of objects in each data set during the annotation process could lead to discrepancies in costs. Some datasets may contain minimal objects per image, while others may have hundreds, making it difficult to determine the cost per image with precision.
Hard to annotate huge bulk of data in a limited timeframe
As Computer Vision and Machine Learning technologies continue to evolve rapidly, the demand for data annotation is skyrocketing. The ability to efficiently and accurately annotate large volumes of data within a limited timeframe has become absolutely critical. Prior to partnering with BasicAI, DataMatica faced significant hurdles in annotating vast amounts of data in fields like autonomous driving. The sheer volume of annotation work and the complexity of requirements demanded significant human resource investment, making it challenging to meet tight deadlines.
Solutions
BasicAI Workflow Management empowers DataMatica to accurately measure the performance of the team
Without an automation platform, it is difficult for DataMatica to accurately track and evaluate the performance by each individual, which can also affect the progress and efficiency of projects.
"To address this challenge," says DataMatica CEO Shih-Hung, "we have harnessed the power of the workflow and performance management system offered by the BasicAI Cloud platform. This has enabled us to successfully move all our projects forward and deliver them with exceptional quality and speed."
BasicAI Cloud Performance Management Interface
In the BasicAI Cloud AI-powered workflow management system, advanced performance stats enable managers to easily monitor task and employee performance, gaining deeper insights into team productivity:
The platform's admin users can easily track the progress of each task and the level of contribution made by each worker.
Additionally, employees can monitor their own individual performance on various tasks.
The ability to export performance data enables the team to view data on all tasks completed by each participant over the past six months, providing a more comprehensive view of team performance.
AI-Powered Annotation Suite boosts DataMatica's labeling efficiency
The demand for AI and machine learning data preparation is expected to continue growing and reach a market size of 1.2 billion US dollars by 2023, according to a report by Cognilytica. As the demand for data labeling increases, improving the speed and efficiency of labeling has become crucial. The DataMatica team has implemented the AI labeling suite on BasicAI Cloud platform, which provides highly efficient and accurate AI automatic annotation through pre-trained models.
"For projects that require significant amounts of data construction in a limited timeframe, particularly in the autonomous driving field, using BasicAI Cloud's automatic labeling technology can complete a large amount of data quickly and cost-effectively, while still achieving a high level of accuracy," said CEO Shih-Hung Chu. "It is fair to say that BasicAI has a top-notch R&D team. We believe that the BasicAI platform has significantly enhanced our efficiency and empowered our team to deliver outstanding results."
Point cloud pre-labeling in BasicAI Cloud Platform
Taking the autonomous driving industry as an example, the BasicAI Cloud system can automate the creation of 2D boxes, 3D boxes, image and point cloud segmentation functions. Before the annotator labels the data, the system can perform model pre-recognition asynchronously, and supports the adjustment of confidence and model inference, maximizing labeling efficiency and accuracy. In particular, efficiency can be improved by 72% in the annotation of 2D & 3D fusion data. At the same time, BasicAI Cloud also supports object tracking in point cloud frame series, eliminating the need for manual frame-by-frame labeling and target linking.
DataMatica's Next Step...
When it comes to the future of the Korean data annotation industry, Shih-Hung said,
"Currently, Korea is leading the way in data annotation in multiple areas, including intelligent CCTV and automated driving control". "Government agencies have allocated billions of Korean won annually for annotation preparation costs. With an increasing demand for new AI services by government agencies, the data annotation market is expected to continue growing."
Data annotation is crucial for AI subfields such as object detection, autonomous driving, and image text recognition. According to Statista, the global third-party data annotation service provider market will reach an impressive $1 billion in 2023, making it a significant area of interest for companies looking to expand globally. Driving this trend, DataMatica has already signed business agreements with top-tier companies such as BasicAI (US) and Siemens (Germany) to capitalize on this trend.
Shih-Hung Chu (KR), CEO, DataMatica
Regarding the partnership with BasicAI, Shih-Hung expressed excitement in the interview:
"BasicAI has garnered global acclaim, including recognition from several Korean companies. At DataMatica, we are thrilled to join hands with exceptional organizations like BasicAI," he said. "We plan to leverage the innovative technology of BasicAI Cloud to promote data and related industries further in Korea. Their data annotation suite has high performance and convenience that is unmatched by other platforms currently available. We believe that its introduction domestically will have a significant impact on our country's technological capabilities."
"Our goal is to leverage the advanced technological capabilities of BasicAI's platform to support Korean companies and industries facing data-related challenges and help them grow their businesses more easily," he said. "We are also excited to explore new technologies that align with the evolving data business landscape in Korea and to collaborate with BasicAI in various ways. Our partnership represents a significant opportunity for us to enhance our technological capabilities and drive the growth of the data industry in Korea," said DataMatica CEO Shih-Hung.
With the collaboration between DataMatica and BasicAI, there is a lot of excitement and anticipation for the future of the Korean data industry.