Scribd
Upload
71 views

AI and Deep Learning Assignment - Stackfusion

The document provides a 2-day task to develop an optical character recognition model and web application. Specifically, it instructs the recipient to: 1. Train an ML model to detect text from images and create a web server allowing users to upload images for OCR inference. 2. Host the web application and share the source code and URL. The submission should also include how data was gathered. 3. Prefer training a neural network for OCR rather than using existing tools. The submission is due in 2 days.

Uploaded by

sk promotions
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
71 views

AI and Deep Learning Assignment - Stackfusion

The document provides a 2-day task to develop an optical character recognition model and web application. Specifically, it instructs the recipient to: 1. Train an ML model to detect text from images and create a web server allowing users to upload images for OCR inference. 2. Host the web application and share the source code and URL. The submission should also include how data was gathered. 3. Prefer training a neural network for OCR rather than using existing tools. The submission is due in 2 days.

Uploaded by

sk promotions
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

AI and Deep Learning

Task Timing: 2 days.


TASK: Train a simple Optical Character Recogniser(OCR) model to identify the text from an image.

1. Train an ML model to detect text from an uploaded image. Find below-attached sample images. Other
images/data can be scraped from the internet.

2. Create a simple web server that hosts a page where a user can upload the image to do an OCR. The web page
should work such that if a user uploads an image, the model trained in Step 1 should inference it. Inference results
should be shown in a simple manner as text output. In case you are familiar with TensorFlow-lite, you can also use
Android-app to demonstrate the model instead of a web server.

Please find the Sample Images below

Address: First Floor, Hartron Complex, Plot-1 Udyog Vihar Phase-1, Gurgaon, Haryana-122016, India
Email: [email protected], Website: www.parkzap.com, Phone: +91-9899337647,+91-9999318399
Address: First Floor, Hartron Complex, Plot-1 Udyog Vihar Phase-1, Gurgaon, Haryana-122016, India
Email: [email protected], Website: www.parkzap.com, Phone: +91-9899337647,+91-9999318399
Only deployed submissions will be accepted
Things to be submit:
You must share with us:

● Share the URL: The application should be hosted/deploy on a URL (use any free cloud server of your
choice e.g. Heroku). Share the link with us once hosted.
● Share the source code: Share the source code for (1)training/inferencing the model (2)Web Server via
your GitHub account
● In Readme mention how you gathered the data

PLEASE NOTE: Prefer training neural networks for the OCR recognition task rather than using available tools like
pytesseract etc.

Email us your submission at [email protected]


Tips:
1. Please use reusable models/tools for image classification.

2. For the web application, use any language of your choice.


The date of submission of the assignment is 2 days from the time the assignment is emailed to you.

Priyanka(+91-9896793165)
Pranay (+91-8076116306)
Note: You have 2 days to complete and submit the assignment. Still, if you will not able to complete kindly get back
to us whatever you have done.
Kindly adhere to the test window, any request to extend the test window will not be entertained.

!!Best of Luck!!

Address: First Floor, Hartron Complex, Plot-1 Udyog Vihar Phase-1, Gurgaon, Haryana-122016, India
Email: [email protected], Website: www.parkzap.com, Phone: +91-9899337647,+91-9999318399

You might also like