Hybrid Frequency Modulation Network for Image Restoration

Authors: Yuning Cui, Mingyu Liu, Wenqi Ren, Alois Knoll

IJCAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments on nine datasets demonstrate that the proposed network achieves state-of-the-art performance for three image restoration tasks, including image dehazing, image defocus deblurring, and image desnowing.
Researcher Affiliation Academia 1Technical University of Munich 2Shenzhen Campus of Sun Yat-sen University {yuning.cui, liumi, knoll}@in.tum.de, renwq3@mail.sysu.edu.cn
Pseudocode No The paper does not contain pseudocode or clearly labeled algorithm blocks, but rather architectural diagrams and mathematical formulations.
Open Source Code Yes The code and models are available at https://github.com/c-yn/CSNet.
Open Datasets Yes We evaluate our model on both daytime and nighttime dehazing datasets. For daytime scenes, the numerical results on a synthetic (SOTS [Li et al., 2018]) and three real-world datasets, i.e., Dense-Haze [Ancuti et al., 2019], NH-HAZE [Ancuti et al., 2020], and O-HAZE [Ancuti et al., 2018]... DPDD [Abuolaim and Brown, 2020] dataset... CSD [Chen et al., 2021] and Snow100K [Liu et al., 2018].
Dataset Splits No The paper mentions training on '256 × 256 patches' and using specific datasets for training and testing, but it does not provide explicit details about the train/validation/test splits of these datasets (e.g., percentages or counts).
Hardware Specification Yes All experiments are carried out on an NVIDIA Tesla A100 GPU.
Software Dependencies No The paper mentions using the 'Adam optimizer' but does not specify its version or the versions of other software libraries or frameworks used.
Experiment Setup Yes Specifically, the model is trained using the Adam optimizer on 256 × 256 patches with a batch size of 8. The initial learning rate is 2e-4, which is gradually reduced to 1e-6 with cosine annealing. We adopt the horizontal flips for data augmentation. According to the complexity of different tasks, we set N (Figure 3 (b)) to 3 for dehazing and desnowing, and 15 for deblurring.