Can I reproduce what you see in tennis tournaments when it comes to modeling the path of balls? With YOLO, Kalman filtering, and other computer vision techniques, it is possible. This is pose es...

Modeling tennis moves/shots (Part 1)

From ETL to building predictive models for multiband signals
In this installment, we describe a bit of my work with multiband signals for seaweed classification. They could be hyperspectral or multiband. I typically work with 10-band signals, where band 10 i...

ETL for multiband data
Overview of AWS pipeline for processing TIF images for downstream analytics. AWS Pipeline for Processing TIF Images (w/ Vegetation Masking) Ingest Amazon S3: Store raw TIF images in ...

Heatmaps using Folium
Folium is a useful library for visualization of maps. Here we see a visualization of the distribution of taxi pickups for a certain day at 5pm in Manhattan. I’ve been using Folium for my research w...

Airflow with stress/strain data
Stress strain data from our experiments are curves. For example, The upstream data are pickled files. They must be processed and uploaded to s3 for downstream consumption. Our Airflow DAGs will...

Optimal transport computation using linear programmming
This is an example of how one can solve OT using linear programming for small cases. The cost matrix is sampled from Uniform(0,1). Naturally, this won’t work well with large $n$, as it’s $O(n^3)$ –...

Image registration
Before any analysis can be done with MRI scans, image registration must be done. Patients can be in many positions when they get their scans. The heads may be slightly tilted or rotated. Image regi...

API gateway
An API gateway serves as an intermediary between the client and the backend services by providing a unified entry point for all API calls. It accepts incoming requests from clients, routes them to ...

Tooling for image segmentation
This is an attempt to do image segmentation using machine learning and FreeSurfer. The brain tumor segmentation (or BraTS) dataset was used. The model is based on a volume-to-volume Generative Adve...

Data-driven investigation of Java memory issue of an actual application (Part 1)
When developing, we seldom see serious memory thrashing issues. I was fortunate enough to see an issue with an application that was built many years ago. (Note: I was not the developer of the app. ...