This is a minimalistic user interface for image quality assessment: the user can express a rank on a list of shown images, one after another.
At the end of the experiment data can be exported for further processing.
Since AFAIK Javascript in a browser sandbox cannot enumerate files in a directory, a different project build is required for each set of images.
The image array is shuffled on each build: if you don't like the generated sequence of images, simply rebuild the project again.
# Install dependencies locally
$ yarn install
# Move your pictures to the images/ folder
# Build the project
$ yarn buildYou can now move dist/ and images/ wherever you want.
Open dist/index.html either locally or on a web server; if everything is OK,
the first image of the randomized set will be shown.
The application is entirely operated by using the keyboard:
| Key | Action |
|---|---|
←, ↑ |
Go to the previous image |
→, ↓ |
Go to the next image |
0 |
Clears the rank given to the current image |
1 : 5 |
Apply a rank to the current image |
s, Enter |
Opens a prompt to save collected as JSON (shift for CSV) |
c, Backspace |
Starts a new session / clears all the collected data |
During the experiment, the given marks are stored in your browser's session storage: reloading the page won't erase them, but closing the window will!
Once exported, you can further process the collected data on your own, for example, by using this MATLAB script to plot the results.
This trivial project is completely free and open source software licensed under the GNU General Public License, version 3.
If you have an idea, found a bug and want to contribute, thanks! Your help is very much appreciated.