Multiview Aggregation for Learning Category-Specific Shape Reconstruction

Authors: Srinath Sridhar, Davis Rempe, Julien Valentin, Bouaziz Sofien, Leonidas J. Guibas

NeurIPS 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments show that our approach is able to produce dense 3D reconstructions of objects that improve in quality as more views are added.
Researcher Affiliation Collaboration Srinath Sridhar1 Davis Rempe1 Julien Valentin2 Sofien Bouaziz2 Leonidas J. Guibas1,3 1Stanford University 2Google Inc. 3Facebook AI Research
Pseudocode No The paper describes its architecture and method through text and figures (e.g., Figure 4 illustrating the network), but it does not include formal pseudocode or algorithm blocks.
Open Source Code No The paper mentions a project website 'geometry.stanford.edu/projects/xnocs' but does not explicitly state that source code for the described methodology is released or provide a direct link to a code repository.
Open Datasets Yes We generated our own dataset, called Shape Net COCO, consisting of object instances from 3 categories commonly used in related work: cars, chairs, and airplanes. We use thousands of instances from the Shape Net [5] repository and additionally augment backgrounds with randomly chosen COCO images [24]. This dataset is harder than previously proposed datasets because of random backgrounds, and widely varying camera distances. To facilitate comparisons with previous work [6, 17], we also generated a simpler dataset, called Shape Net Plain, with white backgrounds and 5 views per object following the camera placement procedure of [17].
Dataset Splits No The paper states, 'We follow the train/test protocol of [36].' However, it does not explicitly provide the percentages or specific counts for training, validation, and test splits within the paper itself.
Hardware Specification No The paper does not explicitly describe the hardware used for running the experiments, such as specific GPU or CPU models.
Software Dependencies No The paper mentions using an 'encoder-decoder architecture based on Seg Net [3]' and the 'Adam optimizer' but does not provide specific version numbers for any software dependencies, libraries, or frameworks.
Experiment Setup Yes Unless otherwise specified, we use a batch size of 1 (multiview) or 2 (single-view), a learning rate of 0.0001, and the Adam optimizer.