Ashtan Mistal's Projects
Tree detection, positioning, and biological family classification from airborne LiDAR data using PointNet++ and mean shift clustering.
A Markdown notetaking application built using Angular and designed using Material Design elements. Built during Winter Break of 2022-23.
A pipeline for semantic segmentation, densification, and planar flattening for improving voxelization and mesh reconstruction quality of airborne LiDAR data.
A proof-of-concept for transforming UBC's Vancouver Campus into a Minecraft world through the use of geospatial data processing and deep learning.
see PHYS_319_Final_Project.pdf.
A Physics-based Introduction to Computer Shading Techniques. Final project for PHYS 420 (Physics Demonstrations). Includes slides and content for a 45 minute presentation geared towards high school students.
Dataloader for UBC campus and corresponding trained PointNet++ model for architectural semantic segmentation on aerial LiDAR data.
A powerful Java-based find and replace application. Performs a sequence of any number of regex find and replace operations on multiple text documents at the click of a button.
Recursive Implementation (via median split) of "Self-Supervised Arbitrary-Scale Point Clouds Upsampling via Implicit Neural Representation (CVPR 2022)"
Personal website.