Case Study: Data Annotation for an AR Video Gaming Company
Customer Profile:
Customer is a gaming company, revolutionizing how people play basketball video games. They are well-known for providing gaming solutions that use Augmented Reality (AR) to allow two or more people in remote areas, play basketball in real-time.
Problem Statement:
To enhance the gaming experience for their gamers, the customer required large amounts of high-quality annotated data to train their AR model. The raw dataset comprised videos of basketball shots, each having 50-80 frames. The customer wanted data labeling to be done for each object in the frames, including backboard, hoop, rim, ball, and make. In-house data labeling of such a large amount of dataset led to – delays, inaccuracy, inefficiency, and increased costs!
Solution:
After analyzing the client’s requirement, the data annotators at Infopine decided to implement the Bounding Boxes video annotation technique. Under this method, a human annotator draws a rectangle around the objects and then adds a label to it.
The implemented data labeling technique resulted in:
- Accurate and fast object detection like balls, rim, hoop, etc.
- Faster training of the AR model
- High-quality annotated data suitable for real-time use
Results:
- Fastened the process of fine-tuning the video games
- Eliminated the need for the customer to hire in-house data annotators
- Reduced the costs involved in the game-building process
- The client could focus more on their core aspects and not worry about getting accurate data to train their model.
Contact Us
So, what are you waiting for? Don’t let your competition get ahead of you. Contact us today to discuss your Data Annotation needs and let us help you achieve your business goals!
