During ISE2023, which is being held at Fira de Barcelona from January 31 to February 2, 2023, ISID, Spanish technology company of AI platforms for audio and video will present the integration of Chat GPT in the video analysis and archiving platform Videoma, to extend the existing documentation functionalities.

In ISID, we are a technology company focused on AI solutions and platforms for advanced video and audio storage and analysis, will be exhibiting its latest generation of the Videoma platform (for advanced analysis and archiving of video, audio and photos) at ISE2023 from January 31 at booth CS620.

In ISID we are exploring various ways to integrate Chat GPT into the video metadata and documentation function currently offered by Videoma. This extracts textual and descriptive information from the images, based on AI modules. With the integration of Chat GPT functionalities, new possibilities open up for extended documentation with greater depth, breadth and interrelatedness than that offered by current AI systems. The different possibilities under study are the following:

  • Summary of trial transcripts. In this case the system would provide summaries of the facts, conclusions from the facts, or relate similar cases for comparative studies.
  • Daily press summary for media. Chat GPT can summarise the information transcribed and extracted from the permanent monitoring of TV stations or streaming.
  • Report of appearances of a public figure. Videoma‘s face detection allows to locate specific characters and the transcript, sent to Chat GPT, can analyse what was said, extract sentiment, etc.
  • Creation of a press release based on statements made by a public figure. In this case, the relational functions of Chat GTP would allow to create press releases automatically, based on statements.
  • Automatic content generation. In the areas of web content or periodic publications, the automatic generation of content, based on a topic that has been detected to be of interest, is vital. In the case of Videoma, Chat GPT could write summaries and descriptions of the metadata that Videoma has extracted.
  • Event summaries or conclusions. Transcription is used as input for Chat GPT in order to summarise or comment on presentations, financial results, interviews, debates, round tables,…
  • Expansion of concepts. Relate several individual concepts (from metadata extracted from video) and elaborate on them.
  • Additional documentation of specific/unknown terms. For those transcripts with very scientific or technical terms, in order to facilitate reading by non-specialists.
  • Structured content synthesis. Chat GPT would be used to synthesise long and complex topics, structuring them logically into coherent sections.
  • Detection of offensive/inappropriate content. By extending Videoma’s intrinsic keyword search functionality, the system could locate content inappropriate for certain age groups and target audiences.

With the integration of new AI modalities, beyond the various algorithms already present in the platform (face detection and recognition, objects, sounds, words, license plates, texts, logos and signs, etc.) Videoma expects to make a qualitative leap in the documentation metadata that the system is capable of generating without human intervention, thus facilitating the cataloging of large libraries or libraries with many new items every day. This facilitates the creation of media libraries that can be easily consulted, substantially reducing the permanent maintenance required to keep them up to date.