A computer vision tool in Python that detects and classifies clusters of nanoparticles in TEM (Transmission Electron Microscopy) images. Built for a researcher who needed automated analysis of microscope imagery at 90%+ detection accuracy.

The tool scans a microscope image, identifies individual nanoparticles, groups them into clusters based on proximity, then analyzes the geometry of each cluster: center point, distances between particles, and angles relative to the cluster center. Clusters are labeled sequentially and classified by shape pattern (e.g. three particles in a line). The output is both a visual overlay on the original image (circles around each cluster with classification annotations) and a text file describing every cluster with its members, geometry, and pattern. Image processing done with scikit-image.

A simple UI with weight sliders lets the researcher tune detection parameters to handle variation between images without touching code.

One-week project from 2015. Small job, but a fun departure from web development into computer vision and scientific tooling.