A Benchmark for Image Retrieval using Distributed Systems

A Benchmark for Image Retrieval using
Distributed Systems over the Internet:
BIRDS-I

aNeil J. Gunther , Giordano Beretta
Imaging Systems Laboratory
HP Laboratories Palo Alto
HPL-2000-162
thDecember 7 , 2000*


Content Based The performance of CBIR algorithms is usually measured on an
Image Retrieval, isolated workstation. In a real-world environment the
client/server, algorithms would only constitute a minor component among the
Internet, many interacting components. The Internet dramatically
performance changes many of the usual assumptions about measuring CBIR
metrics, performance. Any CBIR benchmark should be designed from a
benchmarking, networked systems standpoint. These benchmarks typically
ground truth, introduce communication overhead because the real systems
ontology, they model are distributed applications. We present our
taxonomy, implementation of a client/server benchmark called BIRDS-I to
distributed measure image retrieval performance over the Internet. It has
systems, been designed with the trend toward the use of small
response time personalized wireless systems in mind. Web-based CBIR
implies the use of heterogeneous image sets, imposing certain
constraints on how the images are organized and the type of
performance metrics applicable. BIRDS-I only requires
controlled human intervention for the compilation of the image
collection and none for the generation of ground truth in the
measurement of retrieval accuracy. ...