Introduction

iMerit's data labeling platform is designed to streamline the process of generating high-quality training data to enhance computer vision models. Our latest annotation tool offers a robust set of features aimed at making the labeling process faster and more precise. Key features include a range of annotation options, such as 2D bounding boxes, 3D cuboids, and polylines, allowing for complex object tracking and labeling in both 2D images and 3D point clouds.

Additionally, QuickServe streamlines project setup, allowing users to effortlessly configure and manage labeling projects. It offers real-time project tracking, ensuring seamless collaboration among multiple stakeholders and keeping everyone informed and aligned on project progress.

This comprehensive solution aims to meet the evolving demands of computer vision model training, ensuring efficiency, accuracy, and scalability in data labeling tasks.

This documentation covers the following main topics:

• Account Activation: Activate account on QuickServe

• Project Setup: How to set up a labelling project

• Pre-process Data: How to format and arrange data for import

• Annotation Tool: In-depth guide for learning how to use the labelling tool

• Key Features: Learn how to use powerful tool interactions to speed up the labelling process

• Audit Tool: How to conduct an audit and raise issues observed in labelling submissions

Minimum Requirement

System configuration

• i7 Processor

• 16GB of RAM is the amount of memory for speed and smooth functioning

• 4 GB GPU Card for rendering

Data Related:

• Storage - The dataset must reside in a S3 Bucket. Access pre-defined S3 bucket

• Data Processing - Pre-process LiDAR data