Samridha Shrestha


I specialize in machine learning applications in Computer Vision (CV). I am currently a CV Engineering Applied Scientist at Group42 working on image and video classification, object detection, portrait segmentation, and video summarization. I worked at the NYU MMCV lab to publish a workshop paper at the ECCV ACVR 2020 conference creating a mobile-based crowd detection and evasion technology for the visually impaired. I also contribute to open-source ML EasyOCR and yolov5 projects at GitHub.

My Github Page

Work with the ritpytrading Library


Pip installation available with $ pip install ritpytrading
The ritpytrading python library uses the RIT REST API to interact with the RIT client from the Rotman School of Management at Toronto, to implement trading algorithms.
Full documentation for the pip enabled library available here.

Get Started!

2020 ECCV Conference ACVR Workshop Paper

Our paper on Active Crowd Analysis was accepted at the 2020 ECCV ACVR Workshop. Active Crowd Analysis is a crowd-avoidance system for the visually impaired (VI) that uses images from a VI user’s smartphone camera to relay the risks posed to them based on the proximity, the density, and the motion of crowds. It uses a computationally-efficient neural network to detect crowds and relay the associated risks to the VI user through directional audio alerts.

Check out all the projects here.

Check out now

My Mission To the Open Source Community

Create projects which others can easily follow and extrapolate to meet their own unique ends meet and find amazing solutions to their respective fields.

Easy to Follow

I strive to keep my projects as initiuitive and easy to follow.

Ready to Ship

The Projects are ready to ship, or customizable to your end.

Up to Date

I strive to keep all dependencies fresh.

Collaboration

Anyone who wants to collaborate should contact me!

Contact ME

Have feedback or a project idea you want to collaborate on.

Let's Get In Touch!


Want to reach me through other means, check me out at linkedin, facebook and twitter