Case study in autonomous driving

From：Nexdata Date： 2024-04-02

The customer's mission is to create the mapping solutions for global OEMs and Tier1.

Challenge:
In order to achieve the level of accuracy, they strives for requires multiple different approaches and machine learning models. One of the methods is to identify various special obstacles. Understanding a the obstacles’s precise location relative to the road can be crucial, as it can allow a vehicle to more accurately make the change. They need multiple video data to train and fine-tune the performance of their model. However, annotating individual frames in a video is both time consuming and costly.

Solutions:
Nexdata makes the annotation solution by combining human intelligence with machine learning to improve video annotation speed dramatically. The platform can predict where the annotated object will move in the next frame. Instead of relabeling the entire image from scratch, manual labelers correct annotations as necessary, dragging or resizing persistent labels to fit precisely the object with the annotation.

Results:

Nexdata help the company reduce the cycle time and achieve the level of accuracy efficiently.

Case study in autonomous driving

Recent

Behavior Detection Data: Enhancing Systems through Human Behavior Analysis

Text-to-Speech (TTS) Data: Fueling the Future of Synthetic Voices

Human Voice Datasets: A Key Resource for Speech Technology Development

Previous

Case Study for Speech Data Collection

Next

Why do virtual assistants accurately answer these types of open-ended questions every time?