Getting Started

To launch a project on the iMerit’s 3D Multi-Sensor Fusion Tool, the user will need to cover the following journey.

  1. Define the ontology - The ontology is a comprehensive schema that defines the categories/classes, attributes, relationships, and rules for annotating data.

  2. Create a taskflow - Taskflow is an interactive channel used to manage the execution of tasks. The goal of a taskflow is to ensure that tasks are executed in a defined sequence by the respective expert mapped to the pipeline.

  3. Pre-process Dataset - It is important to organize the data in the tool ingestible format.

  4. Create a job - A job is a comprehensive collection of batches. Each batch consists of a set of tasks assigned to a contributor. A job, hence encapsulates the entirety of work required to complete a larger objective by organizing tasks into manageable groups in the form of batches. Multiple batches can be processed in parallel, and the completion of all batches signifies the completion of a job.

    1. Guide to create a job.

    2. The dataset must reside in an S3 account from where the tool will read the data and present on the tool. Accessing the predefined S3 bucket section will help with this step.

    3. To import the data placed in the S3 bucket in Step 3, go to Import.

  5. To add contributors to the project, the user must share the solution from the Dashboard. The Share section helps the user add more members.

  6. The tool can be accessed on the All Batches UI via clicking on the Launch Task button.

  7. Once the annotations are created, the auditor can audit batches in the Audit tool - How to Audit.

Minimum Configuration

System configuration to support the tool:

  1. i5 Processor

  2. 16GB of RAM is the amount of memory for speed and smooth functioning

  3. 4 GB GPU Card for rendoring

Last updated