Platform has an artificial intelligence, capable of automatically " learn" through the Machine Learning engine, how to identify information that appears within your VOD content and/or images, and transform them into predictions, associated content and metadata.
Before we start, please take a look at the following blog:
Below, we will explain in detail how this module works.
First, it shows the following interface:
First, you can see the filter buttons, which indicate the people who were recognized by the IA and the unknown ones.
In both cases, you will see under each photo the name of the person, or the word Unknown, as shown below:
Identificados
Desconocidos (Uknown)
In the identified window, you will also see a search engine on which you can type a name, the system will show you the characters related to it, click on the one you are looking for and press enter.
If you click on a person's photo, a window with the related information will be shown, it is worth mentioning that not all the fields will always be filled, in this window you will be able to visualize the basic information (1), the name, title (sometimes related to the job position held) and a description of the person, in this section you can delete or rename the person as the case may be.
In the same window will be presented the faces (2) found of the same person within the content on the platform (VOD or Image) and you can edit the name (Not recommended) select one of the faces as default, or simply delete the image, this last option also deletes the content related to that specific image in the AI module.
Finally, you will see the related content (3) where the AI found the person, if you click on one of these, it will open another section of Content Details above this last mentioned, if it is an image, it shows us under the photo a description that usually comes by default or if it was edited from the Image Module, a gallery of related people found in the same and a tab with tags related to the content of the image.
However, if the content is a VOD, it will start playing immediately:
At the bottom of the player, you will see the times in which the person appears in the video clip, if you click on these you will automatically go to that moment on the timeline, in addition you will see some marks referring to these moments.
A la derecha del player, se muestran 7 pestañas con metadata, estas son:
Persons: people found into the same clip, the system shows in green the current one, in blue other people found and in red the uknown.
Keywords: keywords into the video, the number depends on whether or not they are detected by the AI.
Labels: There are usually a lot of labels related to the images contained in the clip, and when you click on any of them, they show you its position within the content.
Sentiments: the AI detects moods and displays them.
Emotions: impressions expressed by the people in the clip.
OCR: character recognition within the video.
Captions: the intelligent system tries to detect the captions said within the video and tries to mark them at the points where they appear, you can edit them as you wish.
If you go back to the filters and select a face shown on the Uknown interface, you will be able to edit its name, by doing so, the person will automatically appear in the identified faces with the characteristics described above.
If you have this module active, you will see inside the VOD content the AI icon.
By clicking on it, the Artificial Intelligence interface will open on the screen, showing the same parameters you would see if you were inside the AI module, and you will be able to visualize each metadata provided directly in the selected clip.
Note: if you delete any record of the identified persons, all the information related to this person within the AI module will be deleted, the system will give you a warning in case you are going to do it, this does NOT remove the associated material inside platform (VOD - Image).
If you have any additional questions about this module and its functions, please feel free to contact us via chat.
Att: Mediastream team