Spherical Image Generation from a Single Image by Considering Scene Symmetry
Authors: Takayuki Hara, Yusuke Mukuta, Tatsuya Harada1513-1521
AAAI 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We conducted experiments to verify the effectiveness of the proposed method. To this end, we used the Sun360 dataset (Xiao et al. 2012), which includes various spherical images, both symmetric and asymmetric, from indoor to outdoor scenes. The data were divided into 50,000 images for training, 10,000 images for testing, and 5,000 images for validation. The spherical image was an RGB image of the equirectangular format with a resolution of 256 512 pixel. |
| Researcher Affiliation | Academia | Takayuki Hara1, Yusuke Mukuta1, 2, Tatsuya Harada1, 2 1The University of Tokyo 2RIKEN |
| Pseudocode | No | The paper describes the network structure and proposed method in text and diagrams, but does not include structured pseudocode or algorithm blocks. |
| Open Source Code | No | The paper states that the source code for a baseline method (PICNet) is publicly available, but it does not provide an explicit statement or link for the source code of the authors' own proposed method. |
| Open Datasets | Yes | To this end, we used the Sun360 dataset (Xiao et al. 2012) |
| Dataset Splits | Yes | The data were divided into 50,000 images for training, 10,000 images for testing, and 5,000 images for validation. |
| Hardware Specification | No | The paper does not explicitly mention the specific hardware (e.g., GPU/CPU models, memory) used for running the experiments. |
| Software Dependencies | No | The paper mentions using the Adam optimizer and implementing functions with full CNNs, but it does not specify software dependencies with version numbers (e.g., Python, PyTorch, TensorFlow versions). |
| Experiment Setup | Yes | We trained the networks from scratch using the Adam optimizer (Kingma and Ba 2014) with a fixed learning rate of 10 4 and a mini-batch size of 8. During the optimization, the weighting factors of the loss function were set as αrec = αgen = 5.0 10 4, βrec = 3.6 10 3 and γ = 0.27; the element of weight vector w corresponding to position v on the unit sphere is w(v) = exp(3 c, v ); and the mixture ratio of the two approximations was set to η = 0.5. Furthermore, we considered C = 5 types of symmetry... |