International: Prime Focus Technologies (PFT) has announced that its AI/ML-powered native media recognition engine CLEARTM Vision Cloud has completed successful POCs (Proof of Concepts) and pilot projects in Media and Entertainment (M&E) industry-specific workflows such as conformance, segmentation, localization, marketing and content QC.
Six months ago, PFT’s Media Services team which has managed more than a million hours of content for global M&E enterprises, deployed CLEAR Vision Cloud for their various client facing projects. The intent was to offer “in-house” learning to CLEAR Vision Cloud’s home-grown AI engines and best-of-breed AI models like AWS, Google and Microsoft. As the machines learned more, the accuracy and actionability of their outputs substantially increased to add significant business value.
“M&E organizations investing in AI/ML should be willing to impart the AI models with learning that is specific to their organization’s unique business needs,” said Muralidhar Sridhar, Vice President, AI/ML Products, PFT. “The machines learn from each and every interaction and improve the accuracy and actionability of their outcomes thereby enhancing the overall efficiency of the process. This is where CLEAR Vision Cloud’s machine wisdom comes into play.”
There are six M&E use cases perfected by PFT for ready deployment. These without exception offer incredible business benefits.
Conformance Use Case: Automatically conform regional edits, subtitles & dub tracks on global masters in a multiple frame rate baseline. CLEAR Vision Cloud ensures very high accuracy across frame rates, enabling a reduction in time, effort, and cost involved in conformance of forced narrations and never before scale in conformance activity.
Segmentation Use Case: Eliminate play-out errors drastically & increase monetization by auto-generating frame-accurate segmentation metadata. With CLEAR Vision Cloud, accurate segment identification is automated guided by visual, audio, and business rules and involves limited manual QC. For short-form workflows, 100% accuracy, with automation in the range of 95-100% and for long-form workflows, increased accuracy and automation that enables 80-90% reduction of cycle time and over 50% in costs have been achieved.
Localization Use Cases: Automatically generate, transcribe, trans-create subtitles in over 60 languages. Apart from auto-generating the captions ensuring a very high level of accuracy, CLEAR Vision Cloud offers side-by-side comparison aiding the caption specialist to fill in the gaps quickly. English language accuracy of 80-90% and 70-85% in other languages is possible with machine learning over a period of time. Automatic subtitle re-timing with 100% accuracy is another use case. CLEAR Vision Cloud’s AI-led comparator can dramatically enhance frame accuracy to near 100% and considerably reduce subtitle re-timing cycle time, leading to better efficiencies and economics.
Marketing Use Case: Locate the right clips needed to build a variety of cross-platform promo material with high-quality discovery of data, automatically. CLEAR Vision Cloud has built one of the best search solutions into the content archive for the global M&E industry. Automatic identification of key shots by leveraging AI is helping save 60-80% of the search time, and with CLEAR Vision Cloud, there is no question of missing any key moments.
Content QC Use Case: Automatically quality-control the bag & tag versions of several promos daily to deliver 100% quality and prevent leakage in your advertising revenue. No more spreadsheets! The AI-powered promo version QC appliance in CLEAR Vision Cloud can drastically reduce manual promo version QC time to seconds as the parameters are auto-detected. The dashboard gives complete visibility into promo status. If a promo asset is rejected for a creative or technical reason, the promo version QC editor can work on it within CLEAR to make the asset good for use.
PFT’s AI/ML offering includes strategic consulting services tailor-made factoring the specific business challenges and unique content of the customer organization. PFT makes AI work for the customer by providing suitable learning to the AI models for specific content and business needs, altering the models to adjust for the specific nuances and conducting experiments to measure the quality of data and sharpening the tools to make accurate and actionable decisions.
