Automating the Counting Process for
State Election Commission
In order to streamline transparency of India's democracy, the Bihar state election commission wanted to automate the counting process to avoid human errors and provide a transparent counting process.
While the manual counting process has been widely used for many years, it is not without its challenges. It is prone to human errors, such as miscounting or misinterpreting ballots, which can lead to incorrect results. Moreover, the process can be time-consuming, particularly in elections with a large number of candidates or high voter turnout.
Our Solution
The EVM Stand
A dedicated stand was designed to securely hold the voting machine. This stand incorporated a camera system that enabled real-time optical character recognition (OCR) technology to accurately interpret and extract information from the ballots.
Real Time OCR
A custom OCR AI model was developed and deployed on the cloud to address the need for accurate digitisation of the EVM's LED display. This model was specifically designed to read and interpret the digits displayed on the EVM in real-time, enabling instant digitization of the displayed information.
votepulseindia.com
In addition to the OCR AI model, a dedicated portal was developed to monitor real-time vote statistics. This portal served as a centralized platform that provided instant access to up-to-date information on the ongoing vote counting process
Key Metrics
This project stands as a remarkable testament to the power of technology in safeguarding the integrity of democracy, ensuring fair representation, and shaping the future of elections.
6,50,000+ training dataset
60,000+ images annotated
80+ hours of model training
4620 EVMs scanned
70000+ datapoints collected
<7s latency to publish results
Cascading AI Pipeline
to achieve 100% accuracy.