Variational Amodal Object Completion

Authors: Huan Ling, David Acuna, Karsten Kreis, Seung Wook Kim, Sanja Fidler

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

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments on complex street scenes demonstrate state-of-the-art performance in amodal mask completion, and showcase high quality scene editing results.
Researcher Affiliation Collaboration NVIDIA1 University of Toronto2 Vector Institute3 {huling ,dacunamarrer, kkreis, seungwookk, sfidler}@nvidia.com
Pseudocode No The paper describes the method using text and equations but does not include structured pseudocode or algorithm blocks.
Open Source Code No The paper does not explicitly state that its source code is publicly available, nor does it provide a link to a code repository.
Open Datasets Yes KINS [27] is a large scale dataset derived from KITTI [10], which contains both instance and amodal annotations. The Cityscapes dataset [7] contains 5,000 images of driving scenes...
Dataset Splits Yes KINS: The dataset consists of 7,474 images for training and 7,517 images for testing. Following [39], we use the first 10% images from the test set as validation set (750 images in total). Cityscapes: ...including 2,975 images for training, 500 for validation, and 1,525 for testing.
Hardware Specification No The paper states 'Please refer to supplementary material for training and model implementation details,' but does not provide specific hardware details in the main text.
Software Dependencies No The paper states 'Please refer to supplementary material for training and model implementation details,' but does not specify software dependencies with version numbers in the main text.
Experiment Setup No The paper states 'Please refer to supplementary material for training and model implementation details,' and mentions optimizing hyperparameters, but does not provide specific values for hyperparameters or system-level training settings in the main text.