EMLight: Lighting Estimation via Spherical Distribution Approximation

Authors: Fangneng Zhan, Changgong Zhang, Yingchen Yu, Yuan Chang, Shijian Lu, Feiying Ma, Xuansong Xie3287-3295

AAAI 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments show that EMLight achieves accurate illumination estimation and the generated relighting in 3D object embedding exhibits superior plausibility and fidelity as compared with state-of-the-art methods.
Researcher Affiliation Collaboration Fangneng Zhan 1, Changgong Zhang 2, Yingchen Yu 1, Yuan Chang 3, Shijian Lu 1, Feiying Ma 2, Xuansong Xie 2 1 Nanyang Technological University 2 DAMO Academy, Alibaba Group 3 Beijing University of Posts and Telecommunications {fnzhan,shijian.lu}@ntu.edu.sg, yingchen001@e.ntu.edu.sg, changyuan@bupt.edu.cn, {changgong.zcg,feiying.mfy}@alibaba-inc.com, xingtong.xxs@taobao.com
Pseudocode No The paper does not contain structured pseudocode or algorithm blocks clearly labeled as "Pseudocode" or "Algorithm".
Open Source Code No The paper does not contain an unambiguous statement that the authors are releasing the code for the work described, nor does it provide a direct link to a source-code repository.
Open Datasets Yes We evaluate EMLight with the Laval Indoor HDR Dataset (Gardner et al. 2017) that consists of 2,100 HDR panoramas taken in a variety of indoor environments.
Dataset Splits No The paper explicitly mentions a test set and training, but does not provide specific details for a separate validation dataset split needed to reproduce the experiment.
Hardware Specification No The paper does not provide specific hardware details such as exact GPU/CPU models, processor types, or memory amounts used for running experiments.
Software Dependencies No The paper mentions "Blender (Hess 2010)" and "Dense Net121" but does not provide specific version numbers for software dependencies like programming languages, frameworks, or libraries used for the implementation or experiments.
Experiment Setup Yes We then employ Vogel s method (Vogel 1979) to generate N (N=128 by default in this work) uniformly distributed anchor points on a unit sphere.