Promising or Elusive? Unsupervised Object Segmentation from Real-world Single Images

Authors: Yafei YANG, Bo Yang

NeurIPS 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We extensively evaluate current unsupervised approaches in a large-scale experimental study. We implement 4 representative methods and train more than 130 models on 6 curated datasets from scratch.
Researcher Affiliation Academia v LAR Group, The Hong Kong Polytechnic University ya-fei.yang@connect.polyu.hk bo.yang@polyu.edu.hk
Pseudocode No The paper does not contain any sections or figures explicitly labeled as 'Pseudocode' or 'Algorithm', nor are there any structured algorithm blocks presented.
Open Source Code Yes The datasets, code and pretrained models are available at https://github.com/vLAR-group/UnsupObjSeg
Open Datasets Yes three commonly-used synthetic datasets: d Sprites [42], Tetris [34] and CLEVR [33], 2) three real-world datasets: YCB [9], Scan Net [17], and COCO [38]
Dataset Splits No Each dataset has about 10000 images for training, 2000 images for testing. No explicit mention of a separate validation split is found.
Hardware Specification No The main body of the paper does not specify the exact hardware used for experiments (e.g., specific GPU models, CPU models, or cloud instances). While the checklist indicates this information is in the Appendix, the Appendix content is not provided in the given text.
Software Dependencies No The main body of the paper does not specify software dependencies with version numbers. While the ethics checklist points to 'Implementation Details in Appendix' for training details, the Appendix content is not provided in the given text.
Experiment Setup No The main body of the paper refers to 'Implementation details' and 'Preparation details for each dataset' being provided in the Appendix, where training details and hyperparameters are typically found. However, the Appendix content is not available in the provided text.