adCenter Labs

Visual product categorization and browsing

adCenterLabs — Sat, 18 Oct 2008 03:54:00 GMT

We have released Visual Product Browsing tool that uses computer vision algorithms to browse and categorize images as a human might, without the need for manual data tagging. Online shoppers could use this technology to visually browse categorized product images and narrow their search to find 'more like this' options. This project addresses the problem of reaching the target audience for advertisers whose products are buried inside a large categorized product database. The product images are retrieved based on attributes such as text and color and arranged into a hierarchical structure for more efficient browsing based on visual similarity.

Please give us your feedback about these tools by contacting us.

Placing non-intrusive ads in videos

adCenterLabs — Sat, 18 Oct 2008 03:36:00 GMT

We have recently updated our website with an exciting tool Intelligent Bug Ads. This tool provides advertisers greater flexibility to serve non-intrusive overlay ads. We have implemented this tool using Microsoft silverlight technology. Try uploading you own logo and watch it appear in the video non-intrusively. There are two options for the advertisers for ads placement: Temporal mode and Spatial mode. The screen is divided into 8 possible ad serving sections. In Temporal Mode, advertisers specify the ad's length and select one of the sections on the screen as target session to serve ads. The system then determines the least intrusive temporal region in which to place ads. In Spatial Mode advertisers specify the ad's length and could still select multiple sections on the screen as target session to serve ads. The system then selects the least intrusive spatial region to place ads.

Please give us your feedback about this tool by contacting us.

Welcome to Microsoft adCenter Labs blog

adCenterLabs — Fri, 27 Jun 2008 00:45:00 GMT

Microsoft is dedicated to pushing the boundaries in digital advertising through continuous innovation in algorithms, systems, and services. To that end, in January 2006, Microsoft adCenter Labs is setup.

Microsoft adCenter Labs is an applied research group dedicated to researching and incubating new digital advertising technologies. It combines the talents of more than 120 top researchers and engineers, all dedicated to developing cutting-edge ways for consumers and advertisers to connect. Our researchers work across a broad range of technology areas, including keyword and content technologies, audience intelligence, ad selection and relevance, social networking, and video and interactive media. The algorithms developed by our team operate in the engines of Microsoft adCenter products such as Paid Search, Content Ads, and Behavioral Targeting. We are also committed to investing in developing online advertising products that will shape the future of advertising through more relevant and meaningful ads across online video, television and mobile devices.

Microsoft adCenter Labs works closely with other research and development labs in Microsoft, and we have built strong collaborations with researchers in Microsoft Research Asia. Researchers at adCenter Labs also actively participate and contribute to academic research communities, attending numerous conferences including SIGKDD, WWW, ICDM, SES, AdTech, and SIGIR. Many of our research results and prototypes were presented at these conferences. Microsoft adCenter Labs is the Platinum Sponsor for KDD 2008, KDD 2007, and KDD 2006, and is the organizer of the International Workshop on Data Mining and Audience Intelligence for Advertising 2008 (ADKDD 2008). We also work closely with the Microsoft adCenter team. Many of the algorithms created at Microsoft adCenter Labs are currently running in the core of Microsoft adCenter products.
Moving forward, our customers and users can expect great accomplishments from adCenter Labs. In the coming years, we hope to release a new set of trend-setting technologies that will become the standard by which advertising products are measured, online and offline.