How do you think the new GigE standards will influence the machine vision industry?
Respond or ask your question now!
Systematics is the science of classifying species and identifying the evolutionary relationships between them. The science is considered a key toward understanding species diversity and helping to alleviate specie extinction.
Until now, to have a specimen accurately identified, it would need to be physically taken to an expert, which could mean traveling to the other side of the world. With the Digital Automated Identification System (DAISY) created by the Natural History Museum (London), which has more than 70 million species in its collection, a user could photograph a specimen with a mobile phone camera, and the identification could be made in seconds by computer.
DAISY uses artificial intelligence and computer-vision technologies to produce virtual collections of authoritatively identified specimens. By sampling electronic images, digitized sounds or digital representations of DNA sequences, DAISY can identify one species from another, including fossilized specimens.
“The DAISY program is based on an unsupervised neural-net generalized pattern recognition algorithm that can accept and work on any sort of digital data,” says Dr. Norm MacLeod, keeper of paleontology at the museum. “Experiments have shown this system can be used to characterize patterns in a wide range of digital image data, even when those data have been subsampled and transformed in various ways to maximize their spatial information content and comparability.”
To date, DAISY has been used to distinguish between training sets of standard RGB TIFF images. Identification accuracies of 85-95% are typical for small-sample trials. However, recent work indicates its performance can be improved by using images of polygon models constructed from 3-D scans of objects, or even raw X, Y, Z coordinate data from those same sorts of scans. Medium and large-sample trials using all three data types are currently in progress.