What the client needs

The company will use artificial intelligence technology to make driving safer; therefore, they need a high-quality data set that provides AI with the necessary information to recognize surroundings and accurately make important road decisions. In particular, because of the need for high precision, customers require a low error rate of only 0.5%.

How we did it

With many years of experience in data annotation in the ADAS field, we quickly set up a team of 35 highly specialized staff to complete 500,000 images for customers in time. In this project, we used data annotation methods such as bounding box, polygon, and polyline to annotate data related to roads, lane boundaries, road surfaces, signs, and traffic cones. Even though we receive high requirements for the quality of annotated images, we still impress customers with our professional competence and high productivity. As a result, completing the project beyond expectations has helped us receive customers’ trust and opened up several new cooperation opportunities in the future.

What the results have

• 500,000 annotated images in 8 months

• Accuracy rate of up to 99.9%

• Data is exported in JSON format