Hierarchical Instance Feature Alignment for 2D Image-Based 3D Shape Retrieval

Authors: Heyu Zhou, Weizhi Nie, Wenhui Li, Dan Song, An-An Liu

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

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments on two popular and novel datasets, MI3DOR and MI3DOR2, validate the superiority of HIFA for 2D imagebased 3D shape retrieval task.
Researcher Affiliation Academia School of Electrical and Information Engineering, Tianjin University, China zhy_std@163.com, {weizhinie, liwenhui, dan.song}@tju.edu.cn, anan0422@gmail.com
Pseudocode No The paper does not contain any structured pseudocode or algorithm blocks.
Open Source Code No The paper does not provide concrete access to source code for the methodology described.
Open Datasets Yes MI3DOR. The MI3DOR dataset [Zhou et al., 2019a] contains 21 categories with 21,000 images and 7,690 3D shapes.
Dataset Splits Yes There are 10,500 images and 3,842 3D shapes for training, while 10,500 images and 3,848 3D shapes for testing.
Hardware Specification No The paper does not provide specific hardware details (e.g., GPU/CPU models, memory) used for running its experiments.
Software Dependencies No The paper mentions using a 'shared 2D CNN' with 'Alex Net' as the backbone and 'Image Net' for pre-training, but does not provide specific software dependencies with version numbers.
Experiment Setup Yes We set the view number as 12 and the camera array setting follows [Su et al., 2015]. We empirically set R as 25 by following [Xu et al., 2019] since the lower value is prone to achieve lower accuracy on target domain while a sufficiently large R may lead to the gradient explosion. We set λ, β as 0.05, 1 and γ = 2 / (1+exp(-10p)) - 1, respectively, where p denotes the training progress varying from 0 to 1.