A selection of research software, interactive systems, and published work in robotics, computer vision, and visual computing.
Simulation tools for robotic grasping, touch, and synthetic data collection.
An Allegro hand simulator for exploring robotic grasps and collecting joint-angle datasets in MuJoCo. It combines primitive-object scenes, interactive finger control, automated grasp trials, and optional tactile-array capture for touch-based shape-recognition research.
View on GitHub
A desktop research framework that converts 3D meshes into labelled synthetic visuotactile data. It brings contact geometry, configurable optical rendering, aligned dataset generation, calibration, and image-classification workflows into one application.
View on GitHubImage and video analysis for understanding real-world scenes.
A vision-based crowd-counting and social-distancing monitoring project using Tiny-YOLOv4 and DeepSORT. The STARPLAN Foot Traffic Monitor demonstration shows the desktop interface for video analysis, people detection, and social-distancing monitoring.
Read the IEEE paperBrowser-based tools for exploring motion and visual data.
An interactive 3D video viewer that turns a clip into a navigable spatiotemporal volume. Width and height form the picture while depth represents time, revealing motion as continuous paths through a single frame stack.
Open Minko