Video has become a dominant method to engage employees and customers through a wide range of use cases. Firms and users alike recognize the immense value in video’s ability to sustain and create a healthy corporate culture. While users push companies for easy access to more tools, smart firms have realized the benefits of a video connected business and take steps to expand their offerings.
However, due to increased offerings enterprises with an extensive media archive often scramble with the problem of converting existing video archives into business value, particularly given the difficulties of content discovery at scale: content categorization is often flawed, and manual tagging is expensive, error-prone and scales poorly.
What are VIDIZMO’s new AI features?
VIDIZMO Enterprise Video Content Management (EVCM) uses Artificial Intelligence (AI) features to draw information from three main pillars audio, visual, and audiovisual content. These contents can be further broken down to OCR, Face Detection and Visual Content Moderation under visual AI features and Voice Activity Detection, Language Identification, Emotion Recognition under audio AI features. Audiovisual AI features are used for Brand Detection, Keyword Extraction, and Topic Modelling.
The shift from Keyword Extraction Model
This method is a step-change from the previous keyword extraction model, which used to pull out and categories only according to explicitly mentioned terms. The new analysis model will help business users extract insight from videos and provide services ranging from face detection to topic modeling; visual content moderation (like deleting obscene visuals) and brands extraction. For facial recognition, it can now automatically recognize over 1 million celebrities – such as actors, actresses, world leaders, athletes, researchers, tech, and business leaders across the globe. All these insights are indexed and presented as structured video data to users.
By indexing these insights allows businesses to make videos more searchable, improve consumer engagement, accessibility of video content, increase the monetizable value for their videos, and content moderation.
VIDIZMO through video indexing enables extraction of understandable, time-stamped metadata and by building search indexes based on that metadata, a more substantial search solution is presented to consumers. For example, indexing faces and spoken words provide the search experience of finding a point in a video where a specific person spoke certain words or when two or more individuals were seen together. Search based on such insights from videos applies to educational institutes, news agencies, entertainment content owners, broadcasters, enterprise LOB apps and particularly to any industry that has a video library that users need to conduct a multidimensional search.
Increasing user engagement is a key business goal for content service providers, as user engagement translates into more revenue opportunities. Video insights unlocked by VIDIZMO’s AI features are indexed to increase user engagement by positioning the relevant video moments to users. As an example, consider an explainer video that explains triangles for the first 2 minutes and buildings in the next 1 minutes. A student reading about buildings will benefit more if the video is positioned starting from the 2-minute marker.
VIDIZMO through indexing extracts metadata for instance spoken words, the text that is part of the video, the people present in a video segment, etc. In 2016, the U.S. Department of Justice revised the Americans with Disabilities Act Title II regulations, introducing the possibility of accessibility requirements. This generated metadata can be used to make content reachable to people with visual or audio disabilities. Also, the player widget provided by VIDIZMO has built-in closed captioning support, and the insights widget can translate the transcript into other languages.
VIDIZMO’s video insights can help increase the monetizable value of videos. As an example, industries that heavily rely on ad revenue (social media, news media, etc.), can deliver more targeted ads by using the extracted insights as additional signals to the ad server, for instance, presenting a rugby shirt (jersey) is more appropriate in the middle of a rugby match vs a tennis match. Through VIDIZMO’s video indexing allows dynamic creation of highlight reels based on snippets from multiple sources. This can assist content owners to supply new content to their syndication network based on what is in-demand by reusing existing assets.
Companies eager to stream their content internationally face a particular problem: A complex networks of national and international regulators who control what is acceptable—and what is not. What the FCC deems a harmless on-screen kiss might be considered as unacceptable by audiences elsewhere around the world. However, with the help of VIDIZMO’s Artificial Intelligence service checks text, and video content for material that is potentially offensive, precarious, or otherwise undesirable. When such material is found, the service applies appropriate labels (flags) to the content which can then be removed this enables companies to ensure their video content meets both local and international regulations as well as ethical standards deemed necessary in each distinct cultures.
For more details on how VIDIZMO can help derive business value with its AI technologies, here are some more links:
Checkout the press release announcing the latest AI features in VIDIZMO.
- VIDIZMO’s Artificial Intelligence bridges the gap between unstructured and structured video data like never before