Active Semi-Supervised Learning Image Validator

This is just an utility script for the validation of images in active semi-supervised learning projects.

The script is written in HTML + JavaScript and therefore it should not require any external installation and/or configuration. You can just download the entire directory and use it in any computer.

The software was tested under . The software was tested on , , and browsers.

You can easily manage the language of buttons and texts setting the LANGUAGE variable in function script: the available languages are only ita and eng up to now.

⚠️ NOTE: The directory of images to validate must be included in the root folder of the project!

Updates

Speech Recognition - Work in Progress

04/05/2023: with the help of Daniele Buschi, the new version of the software can be guided using Speech Recognition. The language settings can be customized using the LANGUAGE variable and the translation file of the project. The only browser which supports this function is Google Chrome and the use of an http server is mandatory. For a ready-to-use application of these new features we have provided the run.sh and run.ps1 scripts for the correct launch of the interface via command line. For their usage the support of Python is required!

Valid/Invalid counters

12/07/2023: adding real-time counters for valid/invalid images to monitor the status of the validation online. This update could be useful to check the effectiveness of the current round of validation and also the rigidity of the validator itself...

Reset current image vote

13/07/2023: adding the reset of the vote for the current image to keep the possibility to express a 3-class validation as valid-invalid-none. This update could be useful to enlarge the possible votes given by the validator (3 is better than 2 😉).