Vahanank

Team Members:

• Oshin S. Wanjari
• Ravikanth Thummula
• Sujith Ampagowni
• Sumit Raj

Problem Statement

• Get automated vechile details like RC Details, Fine History, Owner Details from https://vahan.nic.in/nrservices/faces/user/searchstatus.xhtml
• Should be scalable to scan details for thousands of vehicle numbers

Solution

~The hard one~

• Using open source techstack like Python, Tesseract, Numpy and building a text extraction system

Advantages/Disadvantages:

• Full-control over the system
• No cost to license required
• Difficult to manage
• Need to be SME

Solution

~The Easy one~

• Using AWS Rekognition using their APIs
• Tech Stack: Python, Boto3, Pillow, Jupyter Notebook

Advantages/Disadvantages:

• No servers to manage
• Low-cost
• High productivity and focus on business logic

Future Scope

• Building custom ML models using Tesseract for better OCR recognition
• Using headless browser and run in multi-node server side fashion
• Use database as storage to be a good netizen
• Build ML models to detect change in captcha DOM and minimal code changes