予見眼
Project Active // Yokengan V3.0

Yokengan
"Eye of Foresight" by Akiba-lab.

Autonomous open-architecture smart glasses. An AI assistant that grants you true sight using a local NPU and a projection HUD screen.

The Yokengan Project

Yokengan is a wearable hardware solution created by the Akiba-lab engineering laboratory. Our goal is to create an accessible smart glasses system featuring a real-time HUD (Heads-Up Display).

The product solves problems in the B2B sector (navigation and inspection for logistics and engineers) and opens new horizons in B2C, acting as an uncompromising intelligent assistant for the visually impaired.

Yokengan 3D Render Yokengan 3D Render

Yokengan Evolution

The engineering journey of Akiba-lab. Finding the perfect balance between computing power, autonomy, and ergonomics.

STATUS: REJECTED

Prototype V1

Raspberry Pi Zero

Attempted to run YOLO locally on a microcomputer.

> Result / Analysis:
  • [!] 512MB RAM is too little, constant OOM (Out of Memory) errors.
  • [!] Critical overheating of the board.
  • [!] Unacceptable speed: 1 FPS.
STATUS: REJECTED

Prototype V2

Smartphone + OTG

Using a pocket Android server for computations.

> Result / Analysis:
  • [-] Android limitations on V4L2 drivers (external camera capture).
  • [-] Power limits of the MicroUSB port.
  • [-] System instability.
STATUS: ACTIVE

Final Architecture

NPU Camera Module

Moving heavy machine vision calculations directly into the camera module featuring a built-in neural accelerator.

  • + The system remains cold.
  • + High response speed.
  • + 100% autonomy.

Yokengan Architecture

[ HARDWARE STACK OF THE LABORATORY ]

Vision

Standard Camera Module — capturing the environment in real-time for subsequent object detection.

Display

0.91" / 0.96" OLED Matrix — energy-efficient HUD interface, connected via I2C for text projection.

AI Brain

Google Gemini API Integration via Wi-Fi for instant generation of complex, contextual descriptions of detected objects.

Chassis

Custom frame and enclosures, designed and 3D-printed at Akiba-lab using Bambu Lab and Creality printers.

Yokengan Components Yokengan Components Diagram

How It Works

01

Scan

The NPU-camera scans the environment and identifies the object via hardware processing.

02

Trigger

A physical button on the glasses frame or an auto-timer captures the needed frame and transmits it via a mobile hotspot.

03

Analyze

The Gemini neural network instantly analyzes the data and generates an accurate description (e.g., car brand or sign text).

04

Project

The text is projected directly into the user's eye through an OLED matrix.

HUD View Projection Interface (HUD View)

Biography

Syrym Photo

Hi, I'm Syrym, the founder of Akiba-lab in Kazakhstan. My engineering journey began in childhood, going from looking under car hoods with my dad to competing in Vex Robotics and FTC. But my dream was always bigger. Inspired by anime like Evangelion and Gurren Lagann, I wanted to build real cyberpunk gadgets.

That’s why I started Akiba-lab—to merge anime aesthetics with hardware engineering. Today, I'm developing Yokengan (open-source AR smart glasses) to turn science fiction into accessible tools that enhance human capabilities.