Market Opportunity
NeuralMarker is an accurate Deep Learning models that can work in the real-world
scenarios. NeuralMarker has been built using the expertise of the world-renowned AI experts who have
experience of developing several of production grade AI models.
Problem Statement
To build a CMS to support and manage AI Assisted labelling annotation and training
platform to Speed up Annotations for Object Detection, Segmentation, Keypoints or OCR.
Digiprima had
taken the challenge to build the Single platform for Annotators, Annotation Manager and Data Scientists
In short, a portal was needed to be developed so that customers can
- Upload the images
- Build Automated Active Learning pipelines
- With Built-in AI Models deliver high quality AI models and save 100s of hours
Our Approach – Proposed Solution:
We proposed a solution that provided an efficient way to upload bulk images and annotate them using the drag and drop tool.
- Front End: Vieu.js
- Database: AWS RDS
- Back End: Node JS, Python
- Storage: Amazon S3 bucket
To build an advanced tools to ensure the quality and speed of training data delivery. With this tool Annotators can create complex Annotations using Advanced Tools such as SuperPixel much easily. Following are the key features for Annotators:
- Fast Annotation
- Low Laten-cy
- KeyBoard shortcuts
- Easily Accessible Instructions
- Self Review Annotationss
Challenges- Objectives achieved
- Delivering the system within the stipulated time.
- Understanding the customer requirements was a challenging task at the beginning and we managed to overcome it successfully.
- Providing a user friendly interface with all the complexity was itself a big task on its own which we achieved with our flexible on demand in-house team.
- Technology challenges big data processing
Results – Customer Testimony
We’re still in the early stages of marketing our product and offering it to our customers. It’s been up and running relatively bug-free for about a year. There have been no issues with their work and that’s a good sign. We were able to work with Digiprima Technologies very closely. Our teams got together once a week at least. Digiprima Technologies would show us updates, enhancements they had made, and take requests from us. Their process was on top of things, especially regarding the schedule. I was very happy with her project management skills.
