Online Hashing with Efficient Updating of Binary Codes

Authors: Zhenyu Weng, Yuesheng Zhu12354-12361

AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental The experiments on two multi-label image databases demonstrate the effectiveness and the efficiency of our method for multi-label image retrieval.
Researcher Affiliation Academia Zhenyu Weng, Yuesheng Zhu Communication and Information Security Laboratory, Shenzhen Graduate School, Peking University {wzytumbler, zhuys}@pku.edu.cn
Pseudocode Yes Algorithm 1 OHWEU Input: streaming data (xi, yi), C, K Output: P and R 1: Learn {wk, bk}K k=1 by PCA-ITQ in the initial stage 2: Initialize P0 = [p0 1, ..., p0 K] and R0 = [r0 1, ..., r0 K] 3: for i = 1, 2, ... do 4: Obtain the binary code g i according to Eqn.(5) 5: for j = 1 to K do 6: Update pi j according to Eqn. (13) 7: Update ri j according to Eqn. (28) 8: end for 9: end for
Open Source Code No The paper does not provide any statement about making its source code publicly available or a link to a code repository.
Open Datasets Yes Datasets We compare our method on two commomly-used multilabel image datasets, MS-COCO and NUS-WIDE. (a).MS-COCO dataset (Lin et al. 2014). ... (b). NUS-WIDE dataset (Chua et al. ).
Dataset Splits Yes (a).MS-COCO dataset (Lin et al. 2014). ... We randomly take 4000 images as the queries, and the rest for training and searching. ... (b). NUS-WIDE dataset (Chua et al. ). ... We randomly take 2000 images as the queries, and the rest for training and searching.
Hardware Specification Yes The experiments are run on a computer with CPU I7, 24GB memory. ... The experiments are run on a PC with Intel i7 3.4 GHz CPU, 24 GB memory.
Software Dependencies No The paper does not provide specific version numbers for any software dependencies or libraries used in the experiments.
Experiment Setup Yes As OKH does, we have m = 300 data points to train the hash functions in the initial stage. ... Based on the results, we set C = 0.1 for MS-COCO and NUS-WIDE.