MiSensorKit

Multimodal data collector in your pocket. Capture synchronized RGB, LiDAR depth, 6-DoF camera pose, 3D skeleton, gaze, and sensor data — all in one tap.

MiSensorKit is a multi-modal field recording application designed for iOS devices. It enables synchronized capture of multiple sensor streams within a single recording session, making it the ideal portable tool for building rich multi-modal datasets in the field.

CAPTURE MULTIPLE MODALITIES

RGB frames saved as high-resolution JPEGs (up to 1024px) LiDAR depth maps stored as UInt16 millimeter buffers Full camera intrinsics and 6-DoF extrinsic pose logged every frame as JSON Real-time 3D body skeleton tracking with joint positions via ARKit Front-facing selfie camera capture Gaze estimation and eye tracking data GPS location with latitude, longitude, altitude, speed, heading, and accuracy metrics IMU data including accelerometer, gyroscope, fused attitude (roll/pitch/yaw), and gravity vector Ambient light intensity and color temperature estimates Barometric pressure readings synchronized with every frame FLEXIBLE RECORDING SETUP

Choose which modalities to record, configure sampling frequency from 1 to 60 Hz, and select output resolution from 224 to 1024 px. A built-in preflight check alerts you to missing permissions or hardware limitations before you start. Monitor sensor status, live FPS, stream health, and session size in real time while recording.

BUILT-IN DATA VIEWER

Browse and inspect recorded sessions directly on your device. Scrub through individual frames to view RGB images, depth maps, camera metadata, skeleton data, and sensor readings — without needing to transfer files to a computer.

EXPORT AND SHARE

Export any session as a ZIP archive directly from the app, or share the raw folder via Files and AirDrop. A session summary sheet appears automatically after each recording with key statistics and one-tap sharing options.

STRUCTURED OUTPUT FORMAT

Each recording session creates an organized folder structure with separate directories for RGB, depth, camera pose, skeleton, GPS, IMU, and metadata — ready for your processing pipeline.

REQUIREMENTS

iPhone or iPad with LiDAR sensor (required for depth recording only) iOS 26.0 or later Designed for researchers and developers who need a portable, easy-to-use tool for multi-modal data collection without external hardware or complex calibration setups.