Data annotation tool for deep learning
Demo
You need NVM v12.8.2.
git clone https://github.com/DDoBBy/Annotti.git
npm install
npm start
-
You can select a task between 'Classification' and 'Object Detection'.
You can select multiple folders for your task.
You can add a label by clicking '+' button on the right.
You can change a label color through the popup window that appears when you click the label.
If you check the label, that label will be activated.You can delete label by clicking 'x' button on the label.
Folder : Move to top-level folder
Analytics : Labeling Analytics
Info : Guide for Annotti
Export : Export json file for labeling
Analytics
You can see the graph of labeling when you click the second button on the left.
-
You can annotate images by clicking or dragging images while pushing 'alt' after you activate the label.
You can see full image by clicking thumbnail.
-
You can add a box by dragging while pushing 'alt' after you activate the label.
You can move or resize a box. You can delete a box by pushing 'D'.
results / classification.json
{
"LABELID": {
"name": "LABELNAME",
"files": ["FILENAME", "FILENAME"]
}
}
results / detection-labelInfos.json
{
"LABELID": {
"name": "LABELNAME",
"color": "COLORCODE",
"boxtoFile": {
"BOXID": "FILEID"
},
"numBoxes": "NUMBEROFBOXES"
}
}
results / detection-boxInfos.json
{
"FILEID": {
"BOXID": {
"labelID": "LABELID",
"x1": "X1",
"y1": "Y1",
"x2": "X2",
"y2": "Y2"
}
}
}
results / detection-dataPaths.json
{
"FILEID": "FILEPATH",
"FILEID": "FILEPATH"
}
├─ app
│ ├─ main : Related to main process
│ ├─ project_managers : Related to each project manager
│ ├─ renderers : Related to renderer processes
│ ├─ templates : js files for html
│ ├─ utils
│ ├─ app.global.css
│ ├─ resources
│ │ ├─ imgs : Annotti icons
│ │ └─ gifs : Related to Annotti guide
│ └─ main.js : main process
└─ package.json : npm guide