SIPI Early History

The USC Signal and Image Processing Institute (SIPI) is one of the oldest and most prominent research organizations dedicated to image processing in the world.

Image processing research at USC was initiated in 1962.  In 1972, the Institute was founded by Professors William K. Pratt and Harry C. Andrews with support from the Advanced Research Projects Agency (now DARPA) in the context of developing the ARPANET (the precursor to the modern Internet).  A database of SIPI technical reports by SIPI faculty and students is available that spans from 1965 to the present day.

During the 1970s, research at SIPI focused on the basic theory of image processing with applications to image transmission and communication (e.g., coding and compression), image restoration and enhancement (e.g., denoising and super-resolution), and image analysis (e.g., feature extraction, image understanding, and computer vision). Much of the early work on transform coding (the basis of virtually all audio, image, and video compression standards that enable our modern globally-interconnected world of pervasive digital multimedia) was conducted within SIPI. 

In 1977, SIPI released the first edition of the USC-SIPI Image Database, which provided the world with a collection of high-quality digital images in an era when such images were scarce and image digitization was nontrivial, requiring expensive specialized equipment.  In the time before data sharing was common, this represented a unique public resource that helped democratize the field and enable image processing research around the world. Many images from the SIPI database have become iconic from their decades of widespread use throughout the image processing literature.

As reflected in current faculty interests, SIPI's scope has broadened since the 1970s to encompass the acquisition, enhancement, and understanding of diverse signals and data types across a wide range of applications.  This diversity is enabled by a common core of data science principles and technical foundations that span representation and approximation theory, sampling and interpolation theory, systems theory, estimation theory, statistics, optimization, and learning.  

The tradition of innovation in signal and image processing is a longstanding a pillar of strength within the USC Ming Hsieh Department of Electrical and Computer Engineering.  For decades, SIPI has represented a substantial portion of the research activity within the department, with the diversity and critical mass needed to tackle a broad spectrum of the most pressing signal processing challenges of modern times.