Traco AI uses machine learning technology and neural networks. The result is a system that can save hundreds of hours of media content analysis. This platform acts as a standalone server within an internal domain which is its advantage over other cloud solutions.
The platform can be integrated into other systems via SOAP and REST API. Currently, integration with Avid Interplay Media Central UX and Viz One from Vizrt is developed. This integration provides metadata views and searches as well as tools for analyzing processed data.
While processing media AI Facer automatically recognizes faces based on biometric identification thanks to the 128 level vector. The entire platform works on the principle of eye and nasal vector detection with adjustable minimal face size (min 20 px x 20 pix).
All detected faces are assigned with a unique anonymous ID (identification number / profile). You can then search in which videos and at which point is the person located. At the same time AI Facer allows you to search for groups of people when they are in one shot.
Creation of Database
Each client creates a database individually according to their needs. Based on the recognition each face in the database gets an anonymous ID that the client can name and edit in the administrative environment. The client can also create the database from his own photos, archive footage or with the help of Wikipedia or IMDb. At the same time importing data from the database in LightRoom and Viz Object Store is supported as well. The database stores information about each ID (name, function, and more), TimeStamps, and individual clip names.
Powerful Node System
Thanks to the powerful scalable node system AI Facer can scan files 25 times faster than realtime in maximum of five simultaneous processes.
Integration with MAM
The AI Facer platform can be integrated into any Media Asset Management (Avid MAM, Viz One and others). Therefore it is a suitable solution for any news room, archive or production.
Automatic “Lower Thirds” Generation
The system allows automatic completion of subtitles in broadcast once a person is stored in the database under a particular ID and individual metadata is added. E.g. automatic adding of names and functions in news
Automatic CC Generation
Traco AI provides speech-to-text recognition and automatic closed captioning generation (CC). Internal recognition engine allows offline speech recognition in English, other languages can be recognized through Google Speech API or other third-party systems. It can extract key words from recognized AI text to search for relevant videos. Gained speech metadata can be integrated into other systems like AVID Media Central, See One and many more.
Television production works every day with a high volume of recorded materials and under time pressure. Therefore it is highly desirable for their production process to operate as effective as possible. AI Facer brings an innovative way to streamline production teams work.
For example a broadcast production is set to create a report from the presidential election campaign of all the main candidates. The creators will have to analyze the recorded material for the past few months and find the most important moments of the pre-election meetings and debates. AI Facer will be very useful for this kind of project and its fast processing.
The entire platform is built on flawless face recognition using a 128D biometric mask. AI Facer analyzes the obtained video material and marks all visible faces.
Newly recognized faces that are not yet stored in the database are automatically marked as Unknown Face with an ID number. The user can fill in detailed information for each ID number and then save the entire profile to the database. It is also possible to create a database from photographs and publicly available sources (Wikipedia, IMDB etc.) Profiles can be edited or removed from the database at any time.
The next step is the process of searching. After simply entering keywords, the user will see a list of all the scenes where the seeked person appears. The tool can also find a combination of multiple faces in one shot.
A great addition to AI Facer is the automatic subtitle generating tool – Traco GFX. It takes care of the subtitle generating process with the possibility of attaching predefined graphics. That means the graphics object with the appropriate label will automatically appear together with the person on the scene.
Depending on the client ́s HW, AI Platform works up to 25 times faster than the real time in maximum of five simultaneous processes.
Due to these automated processes, the production teams can easily navigate themselves in the obtained video material and immediately get an overview of the most important scenes.
TRACO AI can be integrated into a wide variety of content management systems and have its functionalities tailored to clients wishes.
TRACO delivers innovative features and capabilities that are useful for sports editors. Our goal is to simplify TV production and help our clients to be as efficient as possible working on their projects.
Editors regularly prepare highlights from the most important sport events for major news. The TRACO AI platform can offer them useful tools that provide instant insight into the obtained video materials and accelerate the process of producing sport reports.
AI Facer analyzes the obtained video material and marks all the visible faces of the athletes. After their recognition process, basic metadata can be assigned with an ID number and then saved to the database. Each saved profile can be easily edited or removed.
Advanced AI functions for example recognize that the score has changed so the editor can initiate ongoing processes in a simple user interface to streamline the production workflow.
TRACO Media Distributor will help you select and process the most important moments of sports matches. It enables automated cropping of short videos that can be instantly shared on the web or social networks. The editorial team gains an effective tool for informing the viewers or advertising the content that will be a part of prime time news.
TRACO GFX was developed for automatic generation of subtitles. When the obtained video materials are analyzed by AI Facer the subtitles can be displayed together with the specific person. GFX can also handle the generation of basic graphics so that the automated process meets the editorial requirements for final broadcasting reports.
Traco AI can be integrated on-premise according to specific customer requirements and the functionality of the entire platform can be customized.
Talent competitions are one of the most popular TV entertainment formats ever. Over the past few years, their popularity has been growing as well as the demands on the production team working on these shows. Traco AI comes with a solution to help editors simplify work on similar projects.
Tens of hours of casting or knockout shots are recorded throughout the shooting. The subsequent orientation in them is often very lengthy and inaccurate. Traco AI brings useful tools to make production of these shows more efficient.
The editors have the task of making collages from all the recordings which illustrates each finalist’s journey through the competition. AI Facer, which can accurately recognize individual faces, is ready to accelerate this job. AI tool function is based on flawless 128D biometric mask which analyzes primarily ophthalmic and nasal vectors.
The speed of analysis is based on the client’s hardware. It can reach a level of speed up to 25 times faster than realtime (with 5 concurrent processes) in pilot projects.
From the beginning of the show, the production can create an database of individual participants with their basic ID information. Photographs or video stills can be used for this purpose.
Each profile in the database contains basic information that can be used to search for individuals. Editors just enter a name and see a list of scenes where the contestant appears. Not even finding a group of people in one scene is an obstacle for Traco AI. This gives the production an absolute overview of the footage.
We have developed Traco AI with the intention of simplifying some lengthy TV production processes. Due to various workflow settings in different types of productions, the platform can be tailored and its functionality modified according to the customer.
The productions are not wasting time searching for important scenes anymore with TRACO AI, therefore working on the final cut is much more efficient.
Working on biographical documents is a great challenge for every production. Manually passing through hundreds of hours of material requires a great deal of patience. Traco AI offers several breakthrough functions to help creators save their time and energy while working on this kind of projects.
Creators started working on biographical movie based on famous music celebrity. Over a period of 50 years long career, countless hours of footage have been recorded. Filmmakers will have to select the most important shots illustrating the main turning points in life.
Given the vast amount of recorded material, it’s quite time-consuming to try to find the most important shots by manual searching.
Traco AI comes up with a revolutionary solution to streamline the time it takes to process a large number of archive footage. AI facer recognizes individuals with its advanced detection and allows basic metadata to be attached to their profiles.
Users can save the analyzed faces into a database, which is followed by searching process for specific people in the video footage. On top of that, users are able to search right from their timeline and a combination of different faces. AI Facer quickly shortlists the times with scenes, where the specific person is located.
For shots taken over several decades, AI Facer can recognize whether it is the same person in different stage of life. Recognition process is based on detailed biometric mask working with the principle of nasal and ocular vectors. Therefore detection accuracy of visible faces works perfectly.
We developed the TRACO AI platform as a Full Rest API, which allows integration to various workplaces. Every production can decide what kind of functions they find useful for their projects or which type of interface they like the most.