Korean Project – Annotation Transportation System

In this case study, let’s find out how we helped address the data annotation needs of a leading Korean automotive company specializing in Advanced Driver Assistance Systems (ADAS).

Client Introduction

Our client, a leader in the Korean automotive industry, is dedicated to advancing driving experiences through innovative ADAS solutions. Seeking to augment their AI capabilities, they partnered with us to obtain meticulously annotated datasets for their systems.

Business Challenge

The client faced the challenge of acquiring accurate annotations for key classes essential for their AI system’s functionality, including Lane/Road boundary, Road Surface, Traffic Sign, Vehicle, and Traffic Cone.

Our Solution

• Specialized annotation techniques: Leveraged diverse annotation techniques such as Bounding Box, Polyline, and Polygon to ensure precise annotations for each class.
• Customized output format: Delivered annotated data in JSON format for seamless integration with the client’s AI systems.

Tech Stack

Achievements

• Quick team setup: Established a dedicated team of 35 members promptly to commence annotation activities.
• Large-scale annotation: Successfully annotated 160,000 images within project timelines.
• Minimal annotation error rate: Achieved an impressively low annotation error rate of 0.5%, ensuring high data accuracy and reliability.