An ELM based local topology preserving hashing

摘要: Hashing learning has become one of the most active research areas in computer vision and multimedia information retrieval with explosively boosted data volume. Mainstream hashing methods adopt a two-stage framework to realize learning. That is, obtain low dimensional embedding encode binary codes respectively. However, this kind divides reduction error encoding loss apart, which is not beneficial preserve original structure. Hence, we propose local topology preserving (LTPH) method reduce simultaneously. To clearly reveal structure, Local Topology Preserving Embedding (LTPE) algorithm proposed paper. LTPE utilizes similarity as well geometry construct topology, can effectively detect Nevertheless, LTPH transductive method, suitable for large scale applications. Considering outstanding global approximation ability fast computation speed Extreme Learning Machine (ELM), an ELM based (ELMLTPH) efficient With facilitation ELM, preserved hamming space. Extensive image experiments are conducted on CIFAR, Caltech 101/256, Corel 10K GIST-1M datasets, demonstrate superiority ELMLTPH compared several state-of-the-art methods.