INTRODUCING EGOSENSE

EgoSense

First-person capture.
Real-world perspective.

A wearable perspective on how people interact with the world. Start with your capture scenario and shape a workflow for embodied data.

EGOSENSEA WEARABLE PERSPECTIVE
Supplied EgoSense product visualization, preserving both camera lenses, housing outline and original wordmark
PRODUCT VISUALIZATION · EGOSENSE
01 / INTRODUCINGFIRST-PERSON · REAL-WORLD · EMBODIED DATAMeet the device ↓

INTERACTIVE ENGINEERING MODEL

Look inside the system.

Original 3D rotation, zoom and explosion with supplied photo textures. Unseen thickness and internals remain engineering approximations; the images do not verify component specifications.

VIEW
EXPLODE 0%
DRAW
See the 8-step assembly story ↓

THE EGOSENSE SYSTEM

One complete capture system

Not just a camera. EgoSense combines a stereo camera module, ready-to-use perception algorithms and a companion app — from wearing to training data in one workflow.

Physical EgoSense cap-mounted device showing the embroidered wordmark and both stereo camera lenses
PHYSICAL DEVICE · CAP-MOUNTED CONFIGURATION
01 · HARDWARE

Wear it like a cap

The stereo camera module on the front (with engraved EGOSENSE logo and rear-only ventilation) and a counterweight battery pod on the back, integrated into an everyday baseball cap.

  • stereo camera module (77.5 mm baseline) + audio + IMU, hardware synced
  • RK3566, 4 GB RAM, 32 GB eMMC + microSD
  • 3+ hours per charge, record while charging
Hand pose and depth estimation
Concept visual · subject to final configuration
02 · ALGORITHM

From raw signals to structured data

Scope a perception pipeline around your task: depth, trajectories, hand poses and aligned exports. Processing and integration are agreed per project.

  • Stereo depth estimation (FastDepth / stereo matching)
  • Visual-inertial SLAM · 6-DoF camera trajectory
  • Hand pose & object detection
  • Time-aligned multi-modal export
EgoSense companion app
Concept visual · subject to final configuration
03 · APP

Control everything from your phone

The EgoSense app connects over Wi-Fi for live stereo preview, one-tap recording and session management.

  • Live stereo preview & device status
  • Start / stop recording, tag sessions
  • Wireless upload & fleet management

02 / DATA APPLICATIONS

Your next breakthrough starts with the right data.

From spatial intelligence to dexterous hands. Define the signals, labels and outputs your team actually needs.

EGOSENSE / SIGNAL EXPLORER01 — 08
Signal schematic · illustrative, not a measured output

Stereo SLAM & 3D reconstruction

Localize the wearer, understand the workspace and validate mapping algorithms.

SLAM · VIO · 6DoF

DATA / INTEGRATION TARGET

Stereo + IMU → 6DoF trajectories, sparse / dense point clouds and 3D reconstruction. Global-shutter capture, hardware sync and accuracy are project configuration requirements to confirm.

Discuss this requirement ↗
HOW IT’S BUILT

How it's built

Scroll to assemble the EgoSense capsule — from bare carrier to the sealed stereo device.

ENGINEERED TO CAPTURE

Wear it. Capture anywhere.

77.5mm

Stereo vision

Synchronized stereo cameras, audio and IMU capture first-person motion and scenes.

32GB

Local storage

eMMC and microSD storage for continuous recording.

2.4 / 5GHz

Wireless

Wi-Fi and Bluetooth for preview, control and data transfer.

≥3h

Mobile capture

Battery-powered recording with USB-C charging.

USE CASES

Different teams. One first-person data foundation.

01

Robotics & embodied AI teams

Build a human demonstration data factory across real homes and industrial spaces. Complement teleoperation with scalable first-person collection and evaluate cross-scene generalization.

Imitation learning · VLA · Dexterity

02

AR / VR & spatial computing

Validate SLAM, first-person perception and gesture interaction on natural movement, hand-object contact and changing environments.

SLAM evaluation · Interaction research

03

Manufacturing & SOP training

Preserve expert know-how, break down assembly and maintenance, and create reviewable footage for quality inspection and training.

Process analysis · Quality · Training

04

Universities & research labs

Build datasets for HCI, computer vision and multimodal learning with explicit calibration, annotation and repeatable evaluation protocols.

HCI · 3D vision · Multimodal datasets

Tell us the task. Let’s define the dataset.

Plan a capture pilot

TECHNICAL SPECIFICATIONS

Technical specifications

The capture foundation. Final configuration and data deliverables are confirmed together.

Form factor
Baseball cap + stereo camera module (124 × 28 × 20 mm) + rear battery pod
Processor
Rockchip RK3566, quad-core Cortex-A55
Memory
4 GB
System storage
32 GB eMMC
Data storage
1× microSD, high-endurance cards supported
Camera
Stereo camera (77.5 mm baseline), USB 2.0 UVC, video / audio / IMU
Wireless
2.4 + 5 GHz Wi-Fi, Bluetooth 5.x
Battery
Single-cell Li-Po, ≥ 3 h continuous recording
Charging
USB-C 5 V, record while charging
Controls
Power button, record button, status LED
Time sync
RTC + monotonic system clock
Thermal
Temperature monitoring, mated to rear ventilation slots
Software
Linux · perception algorithms · iOS / Android app
Debug
USB debug / data port, UART pads

DATA WORKFLOW

A dataset workflow, not just a recording.

  1. 01

    Scope the task

    Agree on scenes, consent, target skills, signals and success criteria.

  2. 02

    Capture in context

    Record stereo, IMU and audio while people perform real tasks.

  3. 03

    Structure the signals

    Calibrate and align streams; scope depth, poses, action labels and privacy processing.

  4. 04

    Validate, then scale

    Review a sample dataset in your training pipeline before fleet deployment.

CONTACT & PILOT

Ask about EgoSense

Tell us your task, target outputs and training stack. Start with a sample-data review and a scoped capture pilot.

1 – 9 units$250per unit
10 – 49 units$220per unit
Best value50+ units$200per unit
Unit price$250
Estimated total$250

Demo inquiry form only — not connected to a backend. No request is sent. Availability, configuration and pricing require confirmation.