People of ACM - Alberto del Bimbo

November 10, 2015

Given current trajectories, what are some exciting potential industrial applications you see arising from image and video analysis, as well as computer vision research in the coming years?

Understanding people's behavior; detecting and classifying facial expressions and emotions from images and sensor data; anticipating the popularity of images or videos that are posted on the web; and interpreting the sentiments expressed in visual content for applications in advertisement and entertainment are only a few of the industrial applications of image and video analysis that will grab the spotlight in the very near future.

At the ACM Multimedia Conference in Brisbane, Australia at the end of October, you participated in a tutorial titled "Image Tag Assignment, Refinement, and Retrieval." How are recent advances in our understanding of image tags changing the field?

With the massive adoption of social network platforms, commenting on digital images has become a common habit. As a result, a huge amount of user-generated images and videos are available with associated tags that describe their content to some extent. This has raised the problem of searching and finding visual data within social platforms but has also revolutionized the scenario of image content understanding. On the one hand, we have to improve social media annotations in order to permit effective retrieval of images and video. On the other hand, we can attempt to learn from images tagged in social contexts and exploit the information provided by tags to understand new visual content. This is different from what computer vision has traditionally done so far. We are just at the beginning of a long path.

As the incoming editor of TOMM, you will have the opportunity to plan special issues around specific topics in the multimedia computing field. What topics would you like to devote a special issue of TOMM to, and why?

Special issues are often overestimated for their capacity to attract readers. Nevertheless they are extremely important if the subject will quickly become popular. This is not easy to predict. Personally, I believe that today media popularity prediction, media emotional content understanding, sensorized reactive environments, and 3D sensor applications are important topics, just to cite a few. But new and very attractive subjects exist in networking and sensors as well.

Multimedia computing is a broad discipline with many potential career pathways in academia or industry. What advice would you give to a younger colleague who is just getting started in this area?

Information technology in this age is all about media computing. Getting aware of technical instruments and scientific achievements is the prerequisite of any future career. But understanding the potentiality and opportunities beyond their current state is the premise for success either in academia or industry. So my advice is "learn and feed your imagination."