Ankit Patel

Ankit Patel

he/him

Engineering AI Solutions for Real-World Impact

Sitemap

A list of all the posts and pages found on the site. For you robots out there, there is an XML version available for digesting as well.

Pages

Posts

Dog Breed Detection with AI šŸ¶šŸ¤–

less than 1 minute read

Published:

Using Convolutional Neural Networks (CNNs), I trained models to classify dog breeds from imagesā€”first from scratch, then using transfer learning with VGG16 and ResNet50. The results? Explore the power of deep learning, transfer learning, and surprising model quirks (like predicting a human as a Black Russian Terrier). Full code available on GitHub.

Breaking Down Europeā€™s Top 5 Soccer Leagues āš½šŸ“Š

less than 1 minute read

Published:

Which league has the most attacking defenders? Which teams defied expectations or crumbled under pressure? Who carried their team single-handedly? Using five seasons of player data (2014ā€“20) and advanced stats like xG, xA, and xGC, I uncover surprising trends in goals, assists, and team dependency. Data meets the beautiful gameā€”dive in! Full code and data available on GitHub.

portfolio

publications

LiDAR-based lane marking extraction through intensity thresholding and deep learning approaches: a pavement-based assessment

Published in The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 2020

This paper improves lane marking extraction across asphalt and concrete using intensity normalization and deep learning.

Recommended citation: Patel, A., et al. (2020). LiDAR-based Lane Marking Extraction through Intensity Thresholding and Deep Learning Approaches: A Pavement-based Assessment. ISPRS Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Download Paper

Intensity thresholding and deep learning based lane marking extraction and lane width estimation from mobile light detection and ranging (LiDAR) point clouds

Published in Remote Sensing, 2021

This paper dives into lane marking extraction using LiDAR intensity normalization and deep learning, achieving higher accuracy for automated road monitoring.

Recommended citation: Patel, A., et al. (2020). Intensity Thresholding and Deep Learning Based Lane Marking Extraction and Lane Width Estimation from Mobile Light Detection and Ranging (LiDAR) Point Clouds. Remote Sensing, 12(9), 1379.
Download Paper

Finding Novel Links in COVID-19 Knowledge Graph Using Graph Embedding Techniques

Published in Smoky Mountains Computational Sciences and Engineering Conference, 2021

This paper leverages graph embeddings and machine learning to predict undiscovered links in a COVID-19 knowledge graph, improving biomedical literature discovery.

Recommended citation: Patel, A., et al. Finding Novel Links in COVID-19 Knowledge Graph Using Graph Embedding Techniques. In Smoky Mountains Computational Sciences and Engineering Conference (pp. 430-441). Cham: Springer International Publishing
Download Paper

talks

teaching

CE203: Principle of Geomatics

Undergraduate course, Purdue University, Fall 2018, Spring 2019 and Fall 2019

Teaching Assistant across three consecutive semesters. Responsibilities included:

  • Delivered lectures during weekly lab session, held office hours and graded weekly lab report submissions
  • Taught students operation of surveying instruments: Total station and GPS
  • Provided instruction on industry-standard software: ArcGIS and PIX4D