Projects

Check out some of my work below!

Black History Footage Colorization Project

Description

Utilizing the U-Net architecture, a popular choice for image segmentation tasks, this model excels in understanding spatial hierarchies and contexts in images, enabling it to add accurate and vibrant colors.

Technologies

Python 3.x, OpenCV, Torch, torchvision, PIL, Numpy, TensorBoard

Check IT OUT

StarTrek WGANGP & SNGAN: Conditional and Controllable Character Generation

Description

In an endeavor to digitally immortalize the captivating characters of "Star Trek: The Next Generation," this project showcases the power of advanced generative models. Leveraging the strengths of Wasserstein Generative Adversarial Networks with Gradient Penalty (WGAN-GP) and Spectral Normalization GAN (SN-GAN) version as well, the model is not only adept at crafting high-fidelity synthetic images of characters from the show but also excels in interpolating between these characters. The project stands out for its conditional generation capabilities, allowing for precise control over the attributes of the generated characters.

Technologies

Python 3.8, PyTorch, NumPy, Matplotlib, torchvision, TensorBoard

Check it out

Facial Tracking with Machine Learning

Description

Tracking, and match move is a huge component of my day to day task. Tracking is a labour-intensive task, I wanted to see if ML can help me generate tracking points automatically without my input.

Technologies

Python, Pytorch, OpenCV, Matplotlib, Numpy.

Check it out!

XFiles Script Generator (USING LSTM)

Description

I’m a huge fan of Xfiles! I wanted to create more stories based on seasons 1-4. With the power of LSTM, I created a character-wise Xfiles Script Generator and Deployed it to Heroku.

Technologies

Python3.8, PyTorch, BeautifulSoup, Flask, Pillow, Numpy

Check IT OUT

Star Trek Character Predictor

Description

This application is a delightful intersection of fandom and technology, specifically designed for Star Trek enthusiasts and machine learning aficionados. If you've ever been curious about which Star Trek character you resemble, this tool is for you. Leveraging the capabilities of a pre-trained VGG-16 neural network model, the application offers an engaging way to find your Star Trek doppelgänger.

Technologies

Python, Matplotlib, Numpy, Pytorch, Torchvision

Check IT OUT!

Neural Network from scratch.

Description

I created an MLP after taking Andrew-Ng course which could be found here. We learned how to build our own deep learning model from scratch, this code is an example of that. I've gotten pretty okay results. With accuracy on the training set being 96% accuracy, and 87% accuracy on the test.

Technologies

Python, Numpy, Imageio, Scipy, Matplotlib.

Check IT OUT!

Ford GoBike System

Description

This project is an Analysis of Lyfts Electric bikes. In this project I go in depth to analyze subscribed users, docked and undocked bikes to find out if it impacts user ride times.

Technologies

Python, Matplotlib, Seaborn Pandas

Check IT OUT!

Estonia Tragedy Exploratory Analysis

Description

The Estonia Tragedy is about a cruise ferry sinking in the Baltic Sea in 1994, I was curious to find out who survived, and if certain categorical features had a play in survival.