This post has been republished via RSS; it originally appeared at: Microsoft Tech Community - Latest Blogs - .
Azure Video Indexer recently released a new feature named "Featured Clothing".
This feature allows publishers and advertisers to leverage deep contextual understanding in-order to place the most relevant ads at the most relevant timing where viewers will be most receptive to engage with the ad.
This insight uses advanced AI to enable a deep understanding of key moments, main characters, important scenes and emotions displayed, all in-order to identify the key clothing items appearing in videos. For these clothing items we provide the exact frames in which they appear with more information (such as bounding boxes and time codes) that can help advertisers and publishers match relevant ads of similar clothing items and place them it in the exact moment where the item is showing within the video.
Imagine you’re watching the NBA finals, it’s the last possession in the last game and your favorite player just scored a threesome that won the championship. Wouldn’t it be great for an ad to pop up at that exact moment with a special sale on that player’s jersey?
Or maybe, you’re watching Carrie Bradshaw rushing to the ferry and losing her Jimmy Choo sandal. There’s a close-up on the item and it looks gorgeous! Why not click on an ad with similar looking items for a special holiday treat for yourself?
Figure 1: showing an illustration of “Featured Clothing” during an NBA game
The global internet advertising market size is expected to surpass the $1T mark by 2027.
This market is dominated by "Direct Response" marketing campaigns. These campaigns are designed to encourage an immediate response from consumers (i.e. clicking , purchasing , subscribing and etc.).
“DR” is heavily based on personalization – tailoring the right ad (at the right time and place) to a specific user (91% of consumers are more likely to shop with brands who recognize, remember, and provide them with personalized offers). This has become a bedrock for effective DR marketing and is broadly used across advanced media platforms and advertisers.
However, recent data collection restrictions are shaking up the industry, redefining how personalization can be used.
The new privacy era
Governments and companies across the globe deploy new regulations and policies designed to enhance consumers’ privacy on the internet.
Under the new policies, companies are required to receive explicit consent from users in-order to track and use their online behavior data (visits to websites, purchases of products and etc.), especially if collected from a third party.
In the early days of advertising – placing an ad in the right medium (based on content) was the only way to ensure effective exposure within a target audience (i.e., an ad for hiking shoes in an outdoor sports magazine).
In a reality where third-party user level data is heavily restricted, contextual advertising, which does not require any usage of personal data, is regaining its place as a prime method to target and optimize ads and is expected to reach to $335.1 Billion by 2026.
Recent developments for in-Video Contextual Advertising
Over the past few years, videos have become an increasingly important part of the marketing and advertising mix.
Video advertising spending in the U.S. was estimated to be $55.34 billion in 2021 and that number is expected to increase by 41.9% by 2024.
However, not long ago the ability to apply contextual advertising was almost exclusively reserved to written text. In video, the level of understanding was relatively basic and usually consisted only of the title, its description, and channel used to distribute it.
Due to recent advancements in artificial intelligence for videos, there has been a dramatic improvement in the ability to provide deep video understanding.
Bring Contextual Advertising Back to stage with Azure Video Indexer
Azure Video Indexer allows for in-video Contextual Ads Placement leveraging the most recent technology advancements in the deep learning domain for video and audio processing. We provide different types of insights, allowing customers to capture what's happening inside their videos and plug in the most relevant ads.
“Featured Clothing” (recently released in public preview) identifies clothing items appearing in videos and ranks them according to their relevancy and importance. The ranking takes into account how dominant the person wearing the clothing is, the scene and the shot the clothing appears in, special audio effects like crowd reaction or laughter and general sentiment and emotions.
This yields a ranked list of bounding boxes in which the clothing items appear with timecodes allowing an indication of what is the best time to show the ad.
Advertisers and publishers can use these insight to apply a matching mechanism - running ads that match the exact clothing item appearing in the video and place it in the most relevant way possible.
“Featured Clothing” is an exciting new way for advertisers and publishers to tap into the most advanced contextual advertising capabilities available today. Achieving best in class ad targeting and relevancy while protecting their users’ privacy.
Capture the most interesting moments combined with the most interesting outfits and offer them to their users. Read more about the AI behind “Featured Clothing” Here.
Join us and share your feedback
For those of you who are new to our technology, we encourage you to get started today with these helpful resources:
- Use Azure Video Indexer website to access product website and get a free trial experience.
- Visit Azure Video Indexer Developer Portal to learn about our APIs
- Search the Azure Video Indexer GitHub repository
- Review our product documentation.
- Get to know the recent features using Azure Video Indexer release notes
- Use Stack overflow community for technical questions.
- To report an issue with Azure Video Indexer (paid account customers) Go to Azure portal Help + support. Create a new support request. Your request will be tracked within SLA.