Regularized Diffusion Process for Visual Retrieval

Authors: Song Bai, Xiang Bai, Qi Tian, Longin Jan Latecki

AAAI 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In this section, we evaluate the validity of the proposed Regularized Diffusion Process (RDP) with toy problems and real retrieval tasks.
Researcher Affiliation Academia Song Bai,1 Xiang Bai,1 Qi Tian,2 Longin Jan Latecki3 1Huazhong University of Science and Technology 2University of Texas at San Antonio, 3Temple University
Pseudocode No No pseudocode or algorithm block is explicitly presented in the paper.
Open Source Code No The paper does not provide an explicit statement or link for open-source code for the methodology.
Open Datasets Yes Following the survey paper (Donoser and Bischof 2013), we assess the effectiveness of the proposed RDP on ORL face dataset, YALE face dataset B (Georghiades, Belhumeur, and Kriegman 2001) and MPEG-7 shape dataset (Latecki, Lak amper, and Eckhardt 2000). We also evaluate the proposed algorithm on the widely-used Ukbench dataset.
Dataset Splits No The paper describes using each image as a query and the rest as the database, but does not provide specific train/validation/test dataset splits with percentages or counts.
Hardware Specification No The paper does not provide specific details about the hardware used to run its experiments.
Software Dependencies No The paper mentions using a 'pre-trained Alex Net' but does not specify software dependencies with version numbers.
Experiment Setup Yes The regularizer μ in Eq. (1) is set to 0.18. ... we set A(1) randomly and the iteration number to 100. ... graph sparsification is applied by only preserving edges within k nearest neighbors. ... For Face and Shape Retrieval, k = 5, k is set to 10. For Natural Image Retrieval, k is set to 4.