The ALCATECH Project
by Jens Arnspang
Within information technology, vision has been studied for decades, perhaps without the feedback to industrial applications as originally expected nor with a full insight into the capabilities of visual exploration of a scene. Common to most efforts has been the study and use of ordinary, commercially available cameras as found commercially available. Within scientific analysis and paradigm building in computational vision the pinhole model has predominantly been used, very much in the same spirit as the Camera Obscura.
Project ALCATECH, whose acronym stands for Alternative Camera Technology, is aimed at studying a series of alternatives to the cameras used hitherto. Breaking existing technical dogmas, inspiration is taken from animal life, where many special purpose eyes (cameras), retinas (sensor chips) and visual tools (optic accessories) are found. There are no standardised eyes in the animal world. On the contrary, different species have surprisingly different vision and display an astonishingly wide variety of eye constructions. Underlying all this variation is the fact that eyes in a hand-in-glove fashion are extremely well adapted to the common tasks a biological creature is involved in, as well as to their environment. Cameras, or more generally optomechanic systems, must also be tailored to fit the task, as well as its operational environment. Nature needs many kinds of eyes and, analogously, there is a need for many kinds of cameras. For a given application or problem domain, a specially tailored optic system, photo sensor design or electronic hardware may convert a hard visual task to an easy, or at least, tractable, problem.
A first ALCATECH Workshop was held at Sjaellands Odde, Denmark, July 1996, hosted by DK SNF and US ONR, organised by Jens Arnspang, DIKU (see communication below) and Ruzena Bajcsy, University of Pennsylvania. The workshop was generously hosted by Ruzena Bajcsy and officially attended and reviewed by Director Bruce Barnes, ONR Washington, through whom the review may be obtained.
The goal of ALCATECH is to use available technological processes (mainly optics, solid state technology and algorithmic processing) to design and implement engineering solutions to specific (yet relevant) visual-based tasks. Central efforts of ALCATECH will be centred on models and implementations of visual capabilities, also borrowing inspiration from the somewhat forgotten techniques of integral photography, invented around the turn of the century. These capabilities are intentionally tailored to a range of visual tasks in order to accelerate a breakthrough in automatic visual inspection and in digital camera technology - for the benefit of both sciences and not least European industry.
Jens Arnspang - University of Copenhagen
Tel: +45 35 32 14 00