Skip to content

Latest commit

 

History

5 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 

Repository files navigation

DRIFT-SENSE

Navigation Error Recovery System

DRIFT-SENSE is an interactive front-end prototype designed to simulate how a navigation error recovery system could detect and correct tiny beam-drift errors during Scanning Electron Microscope (SEM) wafer imaging.

The idea is simple: upload a SEM wafer image, choose a matching algorithm, and let DRIFT-SENSE run through a simulated inspection process. The dashboard shows confidence levels, estimated drift, correction vectors, and a visual heatmap — all inside a real-time instrument-style interface.

Note: DRIFT-SENSE is currently a front-end prototype. There is no backend or real image-processing pipeline. The analysis results are deterministically simulated in the browser using properties of the uploaded file. Your image never leaves the browser.


✨ What You Can Do

🔬 Try Different Matching Algorithms

DRIFT-SENSE provides three simulated matching approaches:

  • NCC — Fast and lightweight, suitable when speed matters.
  • Feature-Based — A balanced approach between speed and accuracy.
  • Hybrid AI — The most accurate simulated method, designed for more challenging images.

Each algorithm has its own characteristics, recommended parameters, and performance profile.

🤖 Automatic Parameter Suggestions

The system automatically suggests parameters based on the uploaded image, including:

  • Patch size
  • Search radius
  • Matching threshold
  • Recommended matching method

The suggestions are generated from properties such as image resolution, compression, and aspect ratio.

📊 Drift Visualization

After a simulation is run, the dashboard visualizes the estimated drift through:

  • Confidence heatmaps
  • Animated bounding boxes
  • Drift direction compass
  • Correction vectors
  • Confidence scores
  • Noise and drift indicators

This makes the simulated inspection process easier to understand at a glance.

🕒 Session History

Every analysis can be saved in the session history.

The dashboard supports up to 16 previous runs, including:

  • Run thumbnails
  • Confidence trends
  • Drift measurements
  • Comparison tables
  • Mini trend charts

This makes it easy to compare different algorithms or images.

🧠 Algorithm Intelligence

DRIFT-SENSE also includes an algorithm guide where you can compare the three matching methods using:

  • Radar charts
  • Benchmark bars
  • Strengths and weaknesses
  • Recommended use cases
  • Parameter guidance

🌍 Multiple Languages

The interface currently supports:

  • English
  • हिन्दी
  • தமிழ்
  • Français

🔍 Zoom & Pan

The SEM image viewer supports mouse-wheel zooming and click-and-drag panning so you can inspect different parts of the image more easily.


🛠️ Tech Stack

DRIFT-SENSE is built entirely on the front end using:

  • React 18 — UI and application logic
  • TypeScript — Type-safe development
  • Vite 6 — Development and production builds
  • Tailwind CSS 4 — Styling
  • Recharts — Radar and benchmark charts
  • Lucide React — Icons
  • HTML Canvas — Heatmap, bounding box, and drift visualization

🚀 Getting Started

Requirements

You'll need:

  • Node.js 18 or newer
  • npm or pnpm

Installation

Clone the project and install the dependencies:

npm i

Start the development server:

npm run dev

Vite will provide a local URL, usually:

http://localhost:5173/

Open that URL in your browser and you're ready to go.


📦 Production Build

To create a production build:

npm run build

The optimized files will be generated inside:

dist/

Because DRIFT-SENSE is completely client-side, the generated dist folder can be deployed to any static hosting platform.


☁️ Deployment

You can deploy the project using platforms such as:

  • Netlify
  • Vercel
  • GitHub Pages
  • Cloudflare Pages
  • Amazon S3 + CloudFront

For platforms that support Vite directly, use:

Build command

npm run build

Output directory

dist

📁 Project Structure

src/
├── app/
│   ├── App.tsx
│   └── components/
│       ├── ui/
│       └── figma/
│
├── imports/
│   └── Background video assets
│
├── styles/
│   └── Global styles and Tailwind entry
│
└── main.tsx

index.html
vite.config.ts

Most of the application logic, simulation behaviour, screens, and dashboard UI are handled inside App.tsx.


🎥 Background Video

The dashboard uses a background video stored inside:

src/imports/

If you replace the video, make sure the filename in the import inside App.tsx exactly matches the file on disk.

File names are case-sensitive, so even a small difference can cause the video to stop loading during the Vite build.


🔐 Privacy

DRIFT-SENSE does not upload your images to a server.

There is currently:

  • No backend
  • No database
  • No external image-processing API
  • No cloud image storage

The simulation runs entirely in the browser.

The analysis values are generated deterministically from the uploaded file's properties, which means the same file will produce consistent simulated results.


⚠️ Prototype Disclaimer

DRIFT-SENSE is a conceptual front-end prototype, not a production SEM navigation or metrology system.

The drift detection, confidence scores, correction vectors, heatmaps, and algorithm benchmarks are simulated for demonstration purposes.

They should not be interpreted as measurements from a real SEM instrument or as actual wafer-inspection results.


About

DriftSense is an AI-powered SEM image analysis system for detecting nanoscale drift in semiconductor wafer imaging. It compares reference and captured images using NCC, feature-based, and hybrid matching to calculate drift vectors, confidence scores, and correction values through an interactive, futuristic analysis dashboard.

Resources

Stars

1 star

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages