Noise2Grad: Extract Image Noise to Denoise
Authors: Huangxing Lin, Yihong Zhuang, Yue Huang, Xinghao Ding, Xiaoqing Liu, Yizhou Yu
IJCAI 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments show that the denoising performance of the proposed method is competitive with other supervised CNNs. |
| Researcher Affiliation | Collaboration | Huangxing Lin1 , Yihong Zhuang1 , Yue Huang1 , Xinghao Ding1 , Xiaoqing Liu2 and Yizhou Yu3 1Xiamen University, China 2Deepwise AI Lab 3The University of Hong Kong |
| Pseudocode | No | The paper does not contain any structured pseudocode or algorithm blocks. |
| Open Source Code | No | We will release our code and datasets soon. |
| Open Datasets | Yes | We use the 4744 natural images in [Ma et al., 2016] to synthesize noisy images (i.e. Dnoise) with software. In addition, 5000 clean images collected from the Internet are adopted as the clean set Dclean. The training dataset is an authorized clinical low-dose CT dataset, which was used for the 2016 NIH-AAPM-Mayo Clinic LDCT Grand Challenge.2 This dataset contains 5946 pairs of images with a slice thickness of 1mm. ... 2https://www.aapm.org/Grand Challenge/Low Dose CT/ |
| Dataset Splits | No | BSD300 [Martin et al., 2001] is the test set of the following experiments. We divide them into 3 parts, 500 pairs as the test set, 2718 pairs as Dnoise, and the remaining 2718 pairs as Dclean. The paper specifies test sets and training data but does not explicitly mention a dedicated validation set with size or purpose for hyperparameter tuning. |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., GPU/CPU models, memory) used for running its experiments, only mentioning the use of 'CNNs' and 'PyTorch'. |
| Software Dependencies | No | The paper mentions 'Py Torch' but does not specify its version number or any other software dependencies with their versions. |
| Experiment Setup | Yes | For simplicity, we adopt a simple U-Net [Ronneberger et al., 2015] as the noise removal module, while the noise approximation module is only composed of one 1 1 convolution layer. We use Py Torch and Adam with a batch size of 1 to train the network. The training images are randomly cropped into 128 128 patches before being input to the network. The learning rate if fixed to 0.0002 for the first 2, 500, 000 iterations and linearly decays to 0 for the next 2, 500, 000 iterations. |