Spatial Computing & AR Hardware2021 – Present (Active Solo Build)Founder & Systems Architect

Augma

Headband-Monocle Augmented Reality Wearable

A bespoke augmented reality wearable adopting a futuristic headband-monocle form factor. Features real-time road sign and obstacle detection using YOLO26, MediaPipe wrist-displacement gesture tracking, an interactive 4-panel glass HUD, and a balanced weight distribution architecture.

MonocleForm Factor
YOLO26Vision Model
Wrist GesturesInteraction
The Problem & Engineering Constraint

The Core Challenge

Traditional dual-lens smart glasses are heavy on the nose bridge (overheating and sliding down), have short battery lives, and suffer from cumbersome input mechanisms that make hands-free computing frustrating.
Technical Architecture & Approach

Engineering Solution & Implementation

Designed an asymmetric headband monocle where battery, compute, and thermal mass sit at the temples and rear headband, while a narrow monocle arm projects visuals directly to one eye. Built an end-to-end Python prototype with YOLO26 neural inference, MediaPipe wrist displacement swipe recognition, point-and-dwell selection, and an interactive HUD compositor.

HARDWARE & SOFTWARE PIPELINE: AUGMA
HEADBAND MONOCLE AR ARCHITECTURE
1. OPTICAL SENSING• Forward-Facing Camera• 1080p Video Stream• IMU Head Tracking• Monocle Eye Path2. DUAL INFERENCE ENGINEYOLO26 Object Detection• Real-time road hazards & signs• Pedestrian & vehicle boundingMediaPipe Hand Tracker• Wrist-displacement swipe (L0)• Point-and-dwell click targetJitter-resistant 8% frame threshold3. 4-PANEL HUD COMPOSITOR▲ UP: Home Status & Time▼ DOWN: Settings & Alpha◀ LEFT: Notification Feed▶ RIGHT: Quick Actions4. MONOCLE• Retinal Path• Low-Weight• Waveguide• Micro-OLEDMASS & THERMAL COUNTERBALANCE DESIGN:Battery (Li-Po) & compute module placed on rear headband and temples. Monocle arm carries minimal front weight, avoiding nasal bridge pressure and slipping.

Figure 4: Augma's end-to-end wearable pipeline combining real-time YOLO26, wrist gesture tracking, and 4-panel HUD.

Headband Monocle Hardware Form FactorAugma Monocle AR Hardware Concept
4-Panel Interactive Glass HUD LayoutAugma 4-Panel Glass HUD Reference
Measured Production Impact

Verified Outcomes & Deliverables

Built working computer vision MVP running real-time YOLO object detection and 4-panel HUD navigation.

Developed wrist-landmark swipe detection algorithm robust to hand jitter and low camera framerates.

Prototyped full industrial CAD and optical pipeline prioritizing ergonomic mass balance over traditional bulky glass frames.

Technologies & Components

System Tooling & Technologies

PythonYOLO26MediaPipeOpenCVEmbedded LinuxOptics & CADSpatial HUD