Lumen | Undergraduate Thesis Project [2016-2017]
A two-part driver assistance application - a distraction detector to notify
a driver when they aren't focusing on the road, and a navigation display to
help them get where they need to go with the least distraction possible. The
distraction detector is a pre-trained neural network deployed on a
RaspberryPi 3 equipped with a Pi Camera that monitors the driver through a
live video feed. The navigation is a web application that projects
directions onto the windshield for the driver to follow.
Developed by a team of 4.
Synchro City [2017]
Embedded Systems Project: A simulation of cars manuevering in a city,
detecting obstacles, following certain paths that don't conflict with each
other to avoide collision. It won the in-class best embedded systems project
award.
Developed by a team of 4.
Designed a Classifier on top of a pretrained Convolution Neural Network
model, trained it and tested its accuracy which gave the 2nd highest rank in
class with a validation accuracy of 83.3%.
Kaggle Competition: Intel and MovileODT Cancer Screening [2017]
Participated in Kaggle competition with a colleague on developing a machine
learning algorithm that learns to accurately identify a cancerous infection
type based on images to help identifying early stages of infection. Team
[Amr & Shehab]'s last contribution ranked 24th out of 703.
Developed by a team of 2.
A Computer Vision project where me and a colleague developed an application
that tries to find hands/arms using our custom color-space for the skin
tone, mask out the background, draw convex envelope around the object, and
then tracks its path from the center of that object.
Developed by a team of 2.
A peer-to-peer distributed systems that allows users of different machines
in the same network to send encrypted images to each other with a view
counter and access privileges. Its core functionality is the following:
- Add/Delete/Download/Display embedded Images.
-
Grant/Revoke Access to users in the database to view certain images with
a set number of views allowed.
- View Downloaded/Allowed-Access Images' list.
Developed by a team of 4.
A Perl application that takes a word document (.docx) and parses the main
elements to generate the corresponding .HTML and .CSS files!
Developed by a team of 3.
Computer Architecture Project: A Desktop GUI application of a MIPS
pipelined CPU with BTB simulator that reads MIPS code and shows the cycle of
each instruction (Fetch -> Decode -> Execute -> Data Memory -> Write Back)
and the content of registers after each cycle.
Developed by a team of 4.