Twitter Feed
Couldn't connect with Twitter

What are you looking for?

Simply enter your keyword and we will help you find what you need.

Audio Annotation Tool

Data annotation is the first and most important step in any machine learning supervised training pipeline. There exist many annotation tools for image classification, object detection, and segmentation — but there is a lack of available tools for annotating audio files.

The developed audio annotator is a tool for annotating audio files, i.e labeling or adding comments to segments of audio file to make it processable for machine learning.

This audio annotator has two interfaces, one for admin and other for common users. Both users have to login via a common login page and are then redirected to their respective interfaces based on user type linked with the email address provided for login.

The user can add his/her annotation groups and start annotating. The annotations are divided into two types, Time-based and word-based annotations.

In “Time-based” annotation, the user has to upload a file and select the audio segment’s start, end intervals and variations manually. After adding all the variations, the user can save them and can download the variations as a “.json” file.

 

 

In “Word-Based” annotation, the user has to upload a file and it will be annotated automatically based on words along with their start and end time using state-of-the-art ML-based speech to text API, later the user has to add variations. Further, a user can edit a wrongly interpreted word and change its start/end time as well. After adding all the variations and making changes to previous ones, the user can save them and can download the variations as a “.json” file.

 

 

The admin can also add his/her own annotation groups and make annotations and download them just like other users. In addition to that, admin can access and download the annotations of other users, admin can also create, view and delete a user.

 

 

Previous Project
Next Project