A Novel Feature Matching Strategy for Large Scale Image Retrieval

Authors: Hao Tang, Hong Liu

IJCAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental The lr SA strategy is evaluated through extensive experiments on five benchmark datasets. Experiments show that the results exceed the retrieval performance of current quantization methods on these datasets. Combined with postprocessing steps, we have achieved competitive results compared with the state-of-the-art methods.
Researcher Affiliation Academia Hao Tang and Hong Liu Engineering Lab on Intelligent Perception for Internet of Things (ELIP) Key Laboratory of Machine Perception (Ministry of Education) Shenzhen Graduate School, Peking University, Beijing 100871, China haotang@sz.pku.edu.cn, hongliu@pku.edu.cn
Pseudocode Yes Algorithm 1 Image retrieval pipeline using lr SA strategy.
Open Source Code No The paper does not provide any explicit statement or link indicating the release of source code for the described methodology.
Open Datasets Yes To evaluate the effectiveness of the proposed lr SA strategy and image retrieval pipeline, we have conducted experiments on five publicly available datasets: Ukbench [Nister and others, 2006], Oxford 5K [Philbin et al., 2007], Paris 6K [Philbin et al., 2008], Holidays [J egou et al., 2008] and MIR Flickr 1M [Huiskes et al., 2010].
Dataset Splits No The paper mentions the use of datasets for evaluation and codebook construction (Flickr60k), but it does not specify explicit train/validation/test splits (e.g., percentages or counts) for the main experiments performed on the benchmark datasets (Oxford, Paris, Holidays, Ukbench).
Hardware Specification No The paper does not provide any specific details about the hardware (e.g., GPU/CPU models, memory) used to conduct the experiments.
Software Dependencies No The paper mentions using 'the FLANN library [Muja and Lowe, 2014] to perform Approximate Nearest Neighbors (ANN) computations' but does not specify a version number for FLANN or any other software dependencies.
Experiment Setup Yes Five parameters are involved in the proposed lr SA strategy: the weight σ, multiple assignment i and T, threshold value t1 and t2. We set i and T the same as [J egou et al., 2010] to 10 and 4, respectively. σ is set to 6250 similar to that in [Philbin et al., 2008]. t1 and t2 are set as 0.6 and 0.4 according to the observation of experimental results, respectively. Two parameters are involved in Hamming Embedding: the Hamming distance threshold and weighting factor β. We set and β to 4 and 7, respectively, the same as those in [Zheng et al., 2014a].