ReyalAI

Solution/Service Overview:

ReyalAI is an end-to-end automatic data annotation tool that enables smart and intelligent data annotation.

Supports high-accuracy, multi-sensor data labeling, making it truly versatile and capable of accelerating computer vision training for autonomous vehicles.

Integration with AWS (better together):

ReyalAI uses AWS S3 to store data.

It has a high level of compatibility with the AWS eco-system components and functionalities facilitating an easy integration.

ReyalAI: Enabling Intelligent Data Labelling for Safe & Secure Drive

Value to AWS:

AWS can provide innovative solutions, solve technical challenges, win deals, and deliver value to our mutual customers.
Expand the reach of AWS into the autonomous vehicle & automotive market.

Value to customers:

Artificial intelligence-assisted annotation
Scalable & modular annotation tool which can be customized as per end user requirement
Advanced canvas for efficient annotations
Compliant with international data security standards like GDPR, CCPA, and SOC2 or DPA

BUSINESS OPPORTUNITIES AND USE CASES

Customer pain points:

Annotation tool’s inability to manage a high volume of data for labelling and lack of customization
Non-availability of an expansive ground truth database that helps in developing & building a robust autonomous vehicle system.
Non-compliance of labelled data with international data security standards like GDPR, CCPA, and SOC2 or DPA
Enable faster time to market while limiting the spend

Gaps filled by ReyalAI:

ReyalAI tool can be customized as per end-user requirements and support high volume labeling.
ReyalAI develops labels that will be used to train and validate autonomous vehicle perception and prediction models
ReyalAI delivers 99.9% annotation accuracy, assisting in the development of a robust autonomous vehicle platform, as well as the reduction in time and cost spent.
GDPR (General Data Protection Regulation) and SoC-compliant solution
The key features of ReyalAI: a) LiDAR Annotations. b) Semantic segmentation - images & videos. c) Polyline, polygon, 2D and 3D annotations.

KEY VERTICALS

AUTOMOTIVE

COMMERCIAL VEHICLES

OUR Engagement model

Target Personas:

Head - Engineering, Autonomous Driving

Director - Engineering, Autonomous Driving

Senior Manager - Engineering, Autonomous Driving

What to Listen:

Absence of labeled data to build a high-quality ground truth dataset to train and validate the platform’s perception and prediction models.
Lack of support in developing different types of labeled data including sensor data (LiDAR) and 2D data.
Minimal support in developing different Annotation including “Polylines, Polygonal Segmentation, Landmark/Key Point Annotation, Bounding Box, Semantic Segmentation, 3D Cuboid Annotation”

Question to Ask:

Type of raw data that needs to be labelled for annotation such as “Images, Videos, Pixel-Level, etc.” .
For what type of labelling data is being processed such as “Object Detection, Lane Detection, Semantic Segmentation, Road Sign Recognition, etc.”.
Are you in need of special use cases including “Driver/Passenger monitoring, auto-parking, blind spot detection, etc.”.

JOINT CUSTOMER SUCCESS

Challenge:

Generating a huge number of labels in various environments including day, night, snow, etc. to train a vehicle perception model to identify objects across different scenarios.
Improve the accuracy of the autonomous vehicle platform perception model by enabling it with a larger annotated/labeled database.

Solution:

To help OEM to have annotations with high quality, Reyal AI’ intelligent data annotation solution/tool ReyalAI was adopted to build and train perception and prediction models
ReyalAI is cloud agnostic and supports high-accuracy multi-sensor data labeling, thus making it versatile and capable of accelerating computer vision training for OEMs’ autonomous vehicle platform
ReyalAI has 99.9% annotation accuracy and is GDPR compliant