Conotation

Local-first image annotation tool with AI-powered segmentation for agriculture. Annotate crop regions, plant parts, and disease areas. Runs entirely on your machine — no cloud, no uploads, no account needed.

🚀 Installation Guide (Step by Step)

Step 1: Install Node.js

You need Node.js installed on your computer. It's a free program that runs Conotation.

1. Go to https://nodejs.org
2. Click the big green button that says "LTS" (recommended)
3. Run the downloaded file and click Next → Next → Install → Finish
4. Restart your computer

To verify it worked: Open Command Prompt (search "cmd" in Start menu) and type:

node --version

You should see something like v20.x.x. If you see an error, restart your computer and try again.

Step 2: Install Conotation

Open Command Prompt and type:

npm install -g conotation

That's it — npm will download and install everything automatically.

Step 3: Run Conotation

In Command Prompt, type:

conotation

Your browser will automatically open at http://localhost:5000

Step 4: Start Annotating

1. Click Open Folder in the top-right
2. Select your folder of crop images
3. Click on any image to start annotating
4. Use SAM tool (press M) to click on a plant/leaf/disease and get an automatic outline
5. Your annotations save automatically

🛑 To Stop the App

Go back to Command Prompt and press Ctrl + C

🔁 Next Time You Want to Use It

Just open Command Prompt and type:

conotation

That's it! No need to install again.

Annotation Workflow

Conotation uses a layered annotation system with three independent tabs:

| Tab | Purpose | Examples | |-----|---------|----------| | Crop Region | Outline the crop area in the image | Healthy Region, Stressed Region, Weed Area | | Plant Part | Identify specific plant parts | Leaf, Stem, Root, Flower, Fruit | | Disease Area | Mark diseased regions | Blight, Rust, Mildew, Leaf Spot, Wilt |

How it works:

• Each tab is independent — annotations from one tab don't interfere with another
• When you switch tabs, other tabs' annotations hide so you see a clean image
• You can annotate the same area on multiple tabs (e.g., mark a leaf as "Plant Part" AND mark disease on it as "Disease Area")
• No annotations are ever deleted when switching between tabs

Tools Available

| Tool | Shortcut | What it does | |------|----------|--------------| | SAM AI | M | Click on object → automatic outline | | Rectangle | R | Drag a box around object | | Polygon | P | Click points to draw a shape | | Brush | B | Paint over an area | | Select | V | Move/resize annotations | | Pan | H | Drag to move around the image | | Point | O | Place a single point marker |

Tool Tips

• SAM AI: Click on an object, the AI generates a mask. Press Enter to accept, Esc to reject. Right-click to mark background points for better results.
• Polygon: Click to add points, double-click or press Enter to finish. Press Esc to cancel. Ctrl+Z removes the last point.
• Brush: Scroll wheel to change brush size. Press E to toggle erase mode.
• Select: Click a shape to select it. Drag vertices to edit. Press Delete to remove.

Quick Labels

After labeling your first annotation, a Quick Label bar appears at the bottom. Click a quick label to auto-assign it to new shapes without the label dialog popping up each time.

Export Your Annotations

Go to the Export tab to download all annotations as a ZIP file in:

• YOLO format (.txt files with normalized coordinates)
• COCO format (single .json file)
• LabelMe format (per-image .json files)

Dataset Splits

Use the Dataset Split panel to divide images into Train/Val/Test sets and download each split separately.

Keyboard Shortcuts

| Key | Action | |-----|--------| | M | SAM tool | | R | Rectangle tool | | P | Polygon tool | | B | Brush tool | | V | Select tool | | H | Pan tool | | O | Point tool | | Ctrl+Z | Undo | | Ctrl+Y | Redo | | Ctrl+S | Save | | Delete | Delete selected shape | | N / → | Next image | | ← | Previous image | | ? | Show shortcuts | | E | Toggle erase (brush mode) | | Enter | Finish polygon / Accept SAM mask | | Esc | Cancel drawing / Reject SAM mask |

Command Line Options

conotation                    # Start on default port 5000
conotation --port=8080        # Use a different port
conotation --dir=C:\my\data   # Use a custom data directory

Troubleshooting

| Problem | Solution | |---------|----------| | 'node' is not recognized | Reinstall Node.js and restart computer | | 'conotation' is not recognized | Run npm install -g conotation again | | Port 5000 already in use | Run conotation --port=8080 instead | | App won't open in browser | Manually go to http://localhost:5000 in Chrome | | Slow on large images | Use Chrome for best performance |

Updating

To update to the latest version:

npm install -g conotation@latest

Requirements

• Windows 10/11, macOS, or Linux
• Node.js 18+ (download from https://nodejs.org)
• Chrome, Firefox, or Edge browser

License

Proprietary — All rights reserved. See LICENSE file.