Towards Domain Invariant Single Image Dehazing

Authors: Pranjay Shyam, Kuk-Jin Yoon, Kyung-Soo Kim9657-9665

AAAI 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We preform extensive experiments to validate the dehazing and domain invariance performance of proposed framework across diverse domains and report state-of-the-art (So TA) results. The source code with pretrained models will be available at https://github.com/PS06/DIDH. Experimental Evaluations Datasets and Evaluation Metrics : In order to evaluate performance of various algorithms across both synthetic and real datasets, exhibiting different haze distributions. We utilize real i.e. NTIRE-18 (Ancuti, Ancuti, and Timofte 2018), NTIRE-19 (Cai et al. 2019), NTIRE-20 (Yuan et al. 2020) and synthetic i.e. SOTS (Li et al. 2019) and Haze-RD (Zhang, Ding, and Sharma 2017) datasets and summarize their properties such as resolution, haze type and average PSNR and SSIM of hazy images in Tab. 2. For evaluating the NTIRE-20 dataset, we first create a subset from training sample and utilize the remaining dataset for training. Furthermore to compare the performance of different So TA algorithms we utilize Peak-Signal-to-Noise Ratio (PSNR) and Structural Similarity Metric (SSIM) as evaluation metrics. Ablation Studies : We examine the effects of different strategies proposed in this paper for improving performance using NTIRE-19 and SOTS-IN datasets. The numerical results for different experiments are summarized in Tab. 5.
Researcher Affiliation Academia Pranjay Shyam1, Kuk-Jin Yoon 2, Kyung-soo Kim 1 1 Mechatronics, Systems and Control Lab 2 Visual Intelligence Lab Department of Mechanical Engineering, Korea Advanced Institute of Science and Technology (KAIST) Daejeon, Republic of Korea, 34141 {pranjayshyam, kjyoon, kyungsookim}@kaist.ac.kr
Pseudocode No The paper does not contain any clearly labeled pseudocode or algorithm blocks.
Open Source Code Yes The source code with pretrained models will be available at https://github.com/PS06/DIDH.
Open Datasets Yes We utilize real i.e. NTIRE-18 (Ancuti, Ancuti, and Timofte 2018), NTIRE-19 (Cai et al. 2019), NTIRE-20 (Yuan et al. 2020) and synthetic i.e. SOTS (Li et al. 2019) and Haze-RD (Zhang, Ding, and Sharma 2017) datasets and summarize their properties such as resolution, haze type and average PSNR and SSIM of hazy images in Tab. 2.
Dataset Splits No The paper mentions creating a subset for evaluation for NTIRE-20 and generally refers to training and evaluating on datasets but does not provide specific, reproducible training, validation, or test split percentages or counts for any dataset.
Hardware Specification Yes For our experiments we utilize a system equipped with Intel 8700-K CPU and 64GB RAM with Nvidia Titan V GPU.
Software Dependencies Yes We design the proposed framework in Pytorch 1.6.
Experiment Setup Yes ADAM (Kingma and Ba 2014) is used as optimizer with β1 = 0.5 and β2 = 0.9 and learning rate of 0.0001 for dehazing and 0.0003 for discriminator networks respectively with a batch size of 4. The input patch is set to square patches of size 512 normalized to [0, 1]. In our experiments we set λ1 = λ2 = 0.5 to balance both LF and HF discriminators. Apart from the aforementioned greedy localized data augmentation (max patch size of 50 50), we also use random horizontal and vertical flipping as additional augmentation techniques.