When words were man’s only indexable form of information storage, there was a lot less information stored. Today, there’s far more information to keep track of—not just text, but speech, images, videos, and sensor data. That means an ever-growing need for signal processing: to store that information more compactly and simplify retrieval, transmit it faster, clarify it, encrypt it, reconstruct it when necessary, and interpret it. And with so many new media and transmission systems constantly being introduced, keeping up can be a challenge.

That’s a major reason why nearly 2000 engineers and academics are expected at the 32nd annual International Conference on Acoustics, Speech, and Signal Processing, to be held from 15 to 20 April in Honolulu. About 1400 papers will be presented, making ICASSP “the world’s largest and most comprehensive technical conference focused on signal processing and its applications,” according to the IEEE Signal Processing Society, the conference sponsor.

Most of the hot topics, says IEEE Fellow Anthony Kuh, the technical program cochair, will be covered in 12 special sessions. Their topics will include signal processing in music, medicine, communications systems, and speech recognition and translation, as well as in sensor networks that rely on distributed processing. The four plenary sessions planned will deal with “The Expanding Utility of the Amazing and Ubiquitous Cellphone,” “50 Years of Progress in Speech Recognition,” “Embedded Intelligence: Beyond Sensor Webs,” and “Network Media Distribution: a Decade of Revolution.”

Sixteen tutorials and two lunchtime panels add to the conference’s diversity. Among their topics will be multimedia and the Internet, the future of signal-processing education, the processing and analysis of music signals, image sensor networks, multimedia search and retrieval, maintaining and enhancing signal quality, and biomedical applications.

ANALYZE ALL Those topics and more to be covered in 72 oral and 96 poster sessions reflect not only technical developments but also social ones—and the ways signal-processing specialties interrelate. For example, as music and multimedia proliferate, the need grows for ways of analyzing and identifying their content so computers can index and retrieve them. This has led content owners to use identification techniques (as well as watermarks and signal fingerprints) to limit or track unauthorized uses.

Such digital rights management overlaps with what’s referred to as forensic processing, which will identify a signal’s source as well as its content, and with processing that is done to recognize what people are saying, who they are (from speech patterns and facial recognition), and even what their gestures and facial expressions signify.

Recognizing gestures and expressions can help with aids for the handicapped, notes IEEE Senior Member and conference cochair Todd Reed. But it also can be used to identify nonverbal threats. That ties in with today’s heightened emphasis on surveillance and security. “To some extent, emotion and gesture recognition would have happened anyway,” Reed says, “but there’s more of it because of 9/11. A lot of funding is now focused in these areas.”

A number of papers will cover wireless sensor networks, which are useful for surveillance but are now used mainly to monitor natural habitats, traffic flow, building movement in high winds, and patients’ vital signs. There are potential military applications, too, says Kuh, that can make use of sensor systems dropped from the air, possibly substituting for human observers on a battlefield.

Biomedical signal processing is also growing in importance. “New medical diagnostic systems generate signals that need to be interpreted and labeled for indexing and retrieval,” says Kuh. Such signals may also benefit from machine interpretation, adds Reed, because, thanks to lower instrument costs, the number of imaging sites is increasing faster than the number of people qualified to interpret the images. With computer help, “you can have less qualified practitioners give diagnoses,” he says.

Machines aren’t the only processors of information, however. So the conference will also cover hearing and psychoacoustics, acoustics, and localization, to help people understand and improve the way they receive and process information.

For further details, go to http://www.ICASSP2007.org.