Training computers to classify pictures and videos

Oct 13, 2010

Spanish researchers have developed a new computer technique that allows to "train" computers to interpret the visual contents of a video or picture. This advance will allow to classify automatically pictures basing on whether individuals or specific objects are present in such images. Videos can also be classified according to specific poses.

At present, computer search and classification of images is made basing on the name of the file, folder or on features as date or size, but the contained was never used for classification purposes. This study conducted by the University of Granada will allow to employ this parameter in the short term for classifying videos according to actions performed by individuals.

The research conducted by Manuel Jesús Marín Jiménez, who is currently working at the University of Cordoba, and coordinated by Professor Nicolás Pérez de la Blanca Capilla, Department of Computing and Artificial Intelligence, University of Granada.

A Specific Pose

Apart from detecting individuals in TV /film shots, this new technique allows to estimate the position of upper limbs (head, chest, arms and forearms) and the automatic classification of video scenes where people appear in a specific pose. Human actions such as walking, jumping, bending down, etc. can also be detected in video sequences.

As Marín Jiménez explains, currently, there is great interest in important international companies as Microsoft or Google in making computers interpret automatically the visual contents of images and video. "Our work is a contribution towards progress in this demanding challenge" the researcher states.

The results of this research have been presented in a number of international conferences such as the International Conference in Pattern Recognition (2006), and the conference on Computer Vision and Pattern Recognition (2008 and 2009).

Example Images:

- People detection on TV:
http://www.robots.ox.ac.uk/~vgg/software/UpperBody/index.html

- Estimation and recognition of human poses:
http://www.robots.ox.ac.uk/~vgg/research/pose_estimation/index.html

Explore further: A social-network illusion that makes things appear more popular than they are

Related Stories

Analyzing audio-visual content

Aug 26, 2010

Filmmakers can identify faults during production with an automatic fault identification and quality assessment system for video sequences. Scientists are also presenting automatic classification for videos at the International ...

Picture this - automatic image categorisation

May 03, 2005

Creating, storing and transmitting visual images has become increasingly easy. Yet the same problem always arises – how to categorise or classify visual images automatically without using external metadata or image thumbnails? ...

Smart computer learns from video

Jun 23, 2010

Swiss researchers have written a computer programme that is able to analyse temporal and spatial patterns of moving objects, and on top of that is capable of learning. This would be a significant aid in traffic ...

Recommended for you

EU open source software project receives green light

Jul 01, 2015

An open source software project involving the University of Southampton to extend the capacity of computational mathematics and interactive computing environments has received over seven million euros in EU funding.

Can computers be creative?

Jul 01, 2015

The EU-funded 'What-if Machine' (WHIM) project not only generates fictional storylines but also judges their potential usefulness and appeal. It represents a major advance in the field of computational creativity.

Algorithm detects nudity in images, offers demo page

Jul 01, 2015

An algorithm has been designed to tell if somebody in a color photo is naked. Isitnude.com launched earlier this month; its demo page invites you to try it out to test its power in nudity detection. You ...

User comments : 0

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.