M-Best-Diverse Labelings for Submodular Energies and Beyond

Authors: Alexander Kirillov, Dmytro Shlezinger, Dmitry P. Vetrov, Carsten Rother, Bogdan Savchynskyy

NeurIPS 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental 4 Experimental evaluation, Quantitative comparison and run-time of the considered methods is provided in Table 1, PASCAL VOC 2012 challenge [9] contains 1449 validation images with known ground truth, which we used for evaluation of diversity methods.
Researcher Affiliation Academia 1 TU Dresden, Dresden, Germany 2 Skoltech, Moscow, Russia
Pseudocode No No clearly labeled pseudocode or algorithm blocks were found in the paper.
Open Source Code No The paper does not provide explicit statements about releasing their source code for the described methodology, nor does it provide a direct link to a code repository.
Open Datasets Yes Category level segmentation on PASCAL VOC 2012 data [9]
Dataset Splits No The PASCAL VOC 2012 challenge [9] contains 1449 validation images with known ground truth, which we used for evaluation of diversity methods. optimally tuned via cross-validation on the validation set in PASCAL VOC 2012. It mentions specific validation data but not explicit splits for train/validation/test from a larger dataset or the standard splits for PASCAL VOC 2012, only that they used the validation set.
Hardware Specification No The paper does not provide specific details about the hardware used for running its experiments (e.g., CPU/GPU models, memory).
Software Dependencies No The paper mentions software components and algorithms like 'min-cut/max-flow problem [21, 28, 16]', 'solver [5]', 'α-expansion [6]', 'α-β-swap [6]' but does not provide specific version numbers for any of them.
Experiment Setup Yes each method was used with the parameter λ (see (2), (4)), optimally tuned via cross-validation.