A Generalized Framework for Edge-Preserving and Structure-Preserving Image Smoothing
Authors: Wei Liu, Pingping Zhang, Yinjie Lei, Xiaolin Huang, Jie Yang, Ian Reid11620-11628
AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | The effectiveness and superior performance of our approach are validated through comprehensive experimental results in a range of applications. Our method is applied to various tasks in the first to the fourth groups to validate the effectiveness. Comparisons with the state-of-the-art approaches in each application are also presented. |
| Researcher Affiliation | Academia | Wei Liu,1,2 Pingping Zhang,3 Yinjie Lei,4 Xiaolin Huang,1,5 Jie Yang,1,5 Ian Reid2 1Department of Automation, Shanghai Jiao Tong University, 2The University of Adelaide 3Dalian University of Technology, 4Sichuan University, 5Institute of Medical Robotics, Shanghai Jiao Tong University |
| Pseudocode | Yes | Algorithm 1 Image Smoothing via Non-convex Nonsmooth Optimization Require: Input image f, guide image g, iteration number N, parameter λ, α, a , b , r , u0 f, with {d, s} 1: for k = 0 : N do 2: With uk, compute ( i,j)k, update (l i,j)k according to Eq. (6) 3: With (l i,j)k, update (μ i,j)k according to Eq. (10) 4: With (l i,j)k and (μ i,j)k, solve for uk+1 according to Eq. (13) (or Eq. (14)) 5: end for Ensure: Smoothed image u N+1 |
| Open Source Code | No | The paper does not explicitly provide a link to open-source code or state that the code is publicly available. |
| Open Datasets | Yes | We test our method on the simulated dateset provided in (Yang et al. 2014). Fig. 9 shows the visual comparison between our result and the results of the recent state-of-the-art approaches. Tab. 1 also shows the quantitative evaluation on the results of different methods. Following the measurement used in (Guo et al. 2018; Li et al. 2016b; Liu et al. 2017a; Yang et al. 2014), the evaluation is measured in terms of mean absolute errors (MAE). We further validate our method on the real data introduced by Ferstl et al. (Ferstl et al. 2013). |
| Dataset Splits | No | The paper mentions using 'simulated To F data' and 'real To F data' but does not provide specific details on how these datasets were split into training, validation, and test sets, such as percentages or sample counts. |
| Hardware Specification | Yes | Our experiments are performed on a PC with an Intel Core i5 3.4GHz CPU (one thread used) and 8GB memory. |
| Software Dependencies | No | The paper mentions 'MATLAB' but does not specify its version number or other software dependencies with version numbers. |
| Experiment Setup | Yes | In all our experiments, we set u0 = f, which is able to produce promising results in each application. Our optimization procedure is summarized in Algorithm 1. ...the parameters are set as ad = ϵ, bd > Im, as = ϵ, bs > Im, rd = rs, α = 0.5, g = f. ...The typical values for rd = rs are 1 3 depending on the texture size. λ is usually smaller than 1. Larger rd, rs, λ can lead larger structures to be removed. The iteration number is set as N = 10. ...We fix rd = rs = 2 and vary λ to control the smoothing strength. λ for the tasks in the first group is usually much larger than that for the ones in the fourth group, for example, the result in Fig. 4(c) is generated with λ = 20. ...We fix α = 0.5, rd = rs, N = 10 for the tasks in both the second and the third groups. We empirically set bd = bs = 0.05Im 0.2Im and rd = rs = 1 5 depending on the applied task and the input noise level. |