Visualizers

Interactive visualizations and tools for exploring various concepts in computer vision, machine learning, and image processing.

Computer Vision & Machine Learning

IoU Playground

An interactive tool for exploring Intersection over Union (IoU) calculations. Drag and move polygons to see how IoU metrics change in real-time, helping build intuition for this fundamental object detection metric.

Pixel Counting vs. Area Estimation

Visualize the difference between counting pixels and estimating areas for segmentation tasks. This tool demonstrates how different sampling strategies affect accuracy when measuring object areas in computer vision applications.

Receptive Field Playground

Explore the concept of receptive fields and effective receptive fields in convolutional neural networks. Interactively adjust network architectures to understand how different layer configurations affect the receptive field size.

Morphology Playground

Interactive demonstration of morphological operations (erosion, dilation, opening, closing) commonly used in image processing. Draw shapes and apply operations to see their effects in real-time.

Polygon Comparison Playground (PoLiS)

Interactive playground for comparing polygons using the PoLiS metric, useful for evaluating building footprint predictions in remote sensing applications.

Image Feature Extraction Demo

Explore image feature extraction techniques and filters interactively. Apply various convolution kernels to understand how feature detectors work in computer vision.

Grid Artifact FFT Playground

Visualize grid artifacts and analyze them using Fast Fourier Transform. Useful for understanding periodic patterns and artifacts in satellite imagery.

Decision Tree Explorer

Interactive demo for exploring decision tree algorithms using the UCI Adult Income Dataset. Build manual trees or use automated algorithms to understand how decision trees classify data.

Random Forest Explorer

Interactive demo showing how random forests improve on single decision trees through ensemble learning. Train hundreds of bootstrapped trees and visualize how majority voting reduces variance and improves accuracy.

Precision / Recall / Prevalence @ Scale

Explore how precision, recall, and false positive counts behave when deploying models at scale with varying prevalence. Understand why a model that looks great on balanced validation sets can produce many false positives in deployment.

Monty Hall Explorer

Interactive playground that lets you play and simulate generalized Monty Hall games with N doors, K host openings, and optional uninformed host behavior to see how switching amplifies your odds.

Utilities

Runner’s Pace Calculator

A practical tool for runners to convert between pace, speed, and race times for various distances. Useful for training planning and race strategy.