Valossa AI April Update Brings State-of-the-Art Face Recognition and Voice Arousal Beta

Our latest cloud system update is now live. Users who have active subscription can start enjoying new features immediately by logging in to Valossa Portal and analysing new videos.

This release contains following features:

More accurate face recognition engine! Over 5000 recognisable celebrities out of the box!

A completely new state of the art face recognition engine for video media use cases. Benefits:

  • Out of the box ability to recognise over 5000 globally known and contemporary celebrities and notable persons
  • More consistent tracking of persons appearing in different parts of the video
  • Begin and end point metadata for each appearance of a person
  • Parallel execution of celebrity recognition with your private face recognition models
  • State-of-the-art recognition accuracy
  • Recognised celebrities contain a wikidata reference with description text to enable intelligent linked knowledge applications on notable people.

The new face recognition has replaced the old face recognition feature. It is available automatically to our subscribers with Standard Metadata, Autopreview, or Face Analysis with Expressions configurations.

Improved private face recognition training!

In parallel to our improved celebrity recognition, you have the option to train your own notable persons directly from your videos faster and more accurately than before!

Use convenient Valossa Reports or your own image uploads to construct your private face recognition gallery. Only a few example face images or training from single video is sufficient to teach the AI for a new identity. Training is also available as a Face Training API.

Emotional arousal from voice – Beta release!

Valossa AI emotional ability is expanding to auditory video modalities. Now the system is also able to measure arousal from human voice.

This feature is in beta and available for clients that are subscribing to Face Analysis with Expressions (also available as an add-on to Standard Metadata).

You can find the Voice Arousal results for your video here:

Valossa Report ->  Emotions and Sentiment tab  ->  Voice Emotion tab

We look forward to hearing your feedback on this powerful functionality!

Your Private Face Recognition Models

We have migrated our paid subscription customers’ private face recognition models automatically to the new, more powerful recognition engine during the update.

For other users interested in trialing or migrating their private face recognition gallery to the new engine, please contact us to learn about your options.

Why wait any longer? Get onboard with the powerful video intelligence tools now!


Latest posts