VoTT Tutorial¶
Tutorial for annotation with VoTT v2.2.0
Visual Object Tagging Tool: An electron app for building end to end Object Detection Models from Images and Videos.
Download VoTT¶
Go to releases page and check the latest version.
Install to macOS¶
export VOTT_VERSION="2.2.0"
wget "https://github.com/microsoft/VoTT/releases/download/v${VOTT_VERSION}/vott-${VOTT_VERSION}-darwin.dmg"
hdiutil mount "vott-${VOTT_VERSION}-darwin.dmg"
cp -r "/Volumes/vott ${VOTT_VERSION}/vott.app" /Applications/
hdiutil detach "/Volumes/vott ${VOTT_VERSION}"
open -a vott.app
Install to Linux¶
export VOTT_VERSION="2.2.0"
wget "https://github.com/microsoft/VoTT/releases/download/v${VOTT_VERSION}/vott-${VOTT_VERSION}-linux.snap"
sudo snap install --dangerous "./vott-${VOTT_VERSION}-linux.snap"
Install to Windows¶
TBW.
Setup a new project¶
You can use sample_project for your tutorial.
The following is a project structure. Put your images in a source
directory.
sample_project
├── source
│ ├── car_1.jpg
│ ├── car_2.jpg
│ └── car_3.jpg
└── target
- Open VoTT app
- Click
New Project
- Fill
Display Name
Source Connection¶
- Click
Add Connection
- Fill
Display Name
Provider
- Select
Local File System
- Select your
source
folder - Click
Save Connection
- Set the connection to
Source Connection
Target Connection¶
- Click
Add Connection
- Fill
Display Name
Provider
- Select
Local File System
- Select your
target
folder - Click
Save Connection
- Set the connection to
Target Connection
Finally, click Save Project
Annotation¶
You can also view the video that shows a manner of annotation (sample.mp4).
Keyboard Shortcuts¶
Shortcut key | Command |
---|---|
W or ArrowUp |
Previous Asset |
S or ArrowDown |
Next Asset |
Ctrl + Arrowkey |
Move Region |
Ctrl + Alt + Arrowkey |
Shrink Region |
Ctrl + Shift + Arrowkey |
Expand Region |
Mouse Controls¶
Mode | Command |
---|---|
Two-point mode | Hold down Ctrl while creating a region |
Square mode | Hold down Shift while creating a region |
Multi-select mode | Hold down Shift while selecting regions |
Exclusive Tracking mode | Ctrl + N to block frame UI allowing a user to create a region on top of existing regions |
Tag ordering¶
Hotkeys of 1 through 0 are assigned to the first ten tags. These can be reordered by using the up/down arrow icons in in the tag editor pane.
Tag locking¶
A tag can be locked for repeated tagging using the lock icon at the top of the tag editor pane.
Tags can also be locked by combining Ctrl or Cmd and the tag hotkey, i.e. Ctrl + 2
would lock the second tag in the list.
Check your progress¶
Export dataset¶
- Go to Export Settings and set your exporting format.
- Click
Save Export Settings
. - Click
Export button
on the tags editor.
The following is a project structure after exporting.
sample_project
├── source
│ ├── car_1.jpg
│ ├── car_2.jpg
│ └── car_3.jpg
└── target
├── 32db62ab992c250ba2312fdc3babc444-asset.json
├── 41e0a9f85a8d20040692c8390317d3ce-asset.json
├── 7ce54d9571515f858d958f3a20cd3ff7-asset.json
├── sample_project-TFRecords-export
│ ├── car_1.tfrecord
│ ├── car_2.tfrecord
│ ├── car_3.tfrecord
│ └── tf_label_map.pbtxt
└── sample_project.vott
Open existing project¶
You can use sample_existing_project as an example for opening local project.
- Click
Open Local Project
- Select VoTT project file (e.g.
sample_project.vott
)