Adaptive Retrieval in a P2P Cloud Datacenter - Multimedia Database Retrieval: Technology and Applications - page 241

Figure 8.22 a, b illustrates retrieval results after three iterations of query modifi-
cation, showing snapshots from the P2P retrieval system. The recall rate at the first
iteration was 46.6 %, whereas at the third iteration, the recall rate was improved to
80.9 %.

8.6 Summary

The chapter presents methods for organizing images and videos in a cloud dat-
acenter. A distributed database can be managed by structured and unstructured
P2P networks. In both scenarios, self-organization plays an important role for
automatic clustering of multimedia objects, and thus, the clustering of nodes for
effective indexing and retrieval. The self-organizing tree map is employed for
the cluster-identification search system, which offers online search by pinpointing
the relevant nodes without traversing all the participating nodes. Furthermore,
the self-organizing tree map and the adaptive cosine network are effective tools
for implementing pseudo-relevance feedback, minimizing bandwidth and user
workload in image and video retrieval processes.