24小时热门版块排行榜

南方科技大学公共卫生及应急管理学院2026级博士研究生招生报考通知（长期有效）

返回列表

【悬赏金币】回答本帖问题，作者wqshmzh将赠送您 3 个金币

当前只显示满足指定条件的回帖，点击这里查看本话题的所有回帖

wqshmzh

铁虫 (小有名气)

应助: 0 (幼儿园)
金币: 147.5
散金: 10
红花: 2
帖子: 67
在线: 37.7小时
虫号: 15899261
注册: 2019-08-10
专业: 模式识别

[求助] 投Neurocomputing，还有没有必要改好了以后重新投回去，求高人指点已有3人参与

以下是原文

please find below the referee reports. based on these and the corresponding recommendations, i have to reach the sad conclusion that your paper

noise image segmentation with adaptive non-local constraint and kl divergence of fuzzy clustering

cannot be accepted for publication in neurocomputing. i hope that the referees' comments and suggestions are nevertheless useful and can help you improving your scientific work and/or presentation.

hereby i would like to thank you for submitting your work to neurocomputing and welcome you to consider us again in the future.

please proceed to the following link to update your personal classifications and keywords, if necessary:
******

sincerely,

editor in chief
neurocomputing

reviewers' comments:

reviewer #1: this manuscript presents a work on developing an improvements to fuzzy c-means clustering.  the improvements are done by incorporated non-local constrain (from non-local means filter), and kullback-leibler (kl) divergence.  good background study has been presented, methodology has been explained with justifications, and the results are impressive, as compared with other methods, for both gray-scale and color image segmentation.
however, for improvements:
1) it would be much better if the authors could define what is nlcklfcm stand for?
2) similarly, it would be better if the authors could define all abbreviations in this manuscript, such as those on page 2.  for example, what is bcfcm stand for?
3) in equation(31), instead of "lg", better to use "log_10" or  "log_2".

reviewer #2: this paper proposes an improved fcm with adaptive non-local constraints and kullback-leibler divergence incorporating spatial information for noise image segmentation. the experimental results demonstrate that the algorithm show better performance than the current state-of-the-art approaches. however, the proposed algorithm can be considered as a combination of some existing algorithms, and the novelty of this manuscript is limited.

in addition, the manuscript is affected by many flaws which limit the correct assessment of its value:

1) section 2.1, what is the specific meaning of gauss euclidean distance? is it gaussian weighted euclidean distance?
2) there are many studies adopting kl divergence to measure the proximity between cluster membership in fuzzy c-means clustering[1][2][3], but no relevant papers have been cited in the paper;
3) the purpose of the nlcklfcm is to solve the problem of segmentation of noisy images. in the experiment, only the simulated noisy images are tested. some experiments on real noise images such as medical images or some public image datasets show to be tested.
4) english writing should be carefully improved, and there are quite a few confusing sentences throughout the manuscript.

minor comments:
there are also some minor errors, for example,
(1) eq.(31)

in conclusion, due to the quality of presentation and the lack of experiments, i think the paper cannot be accepted in its present form. a very major revision may necessary before the re-submission.

[1] chatzis s p. a fuzzy c-means-type algorithm for clustering of data with mixed numeric and categorical attributes employing a probabilistic dissimilarity functional[j]. expert systems with applications, 2011, 38(7): 8684-8689.
[2] zhu h, pan x. robust fuzzy clustering using nonsymmetric student? st finite mixture model for mr image segmentation[j]. neurocomputing, 2016, 175: 500-514.
[3] gharieb r r, gendy g, abdelfattah a, et al. adaptive local data and membership based kl divergence incorporating c-means algorithm for fuzzy image segmentation[j]. applied soft computing, 2017, 59: 143-152.

reviewer #3: the authors propose an improved fcm with adaptive non-local constraints and kullback-leibler divergence incorporating spatial information for noise image segmentation. experiments show that the proposed method achieves much better performance in noise image segmentation.overall, the article is well organized and its presentation is good. in section 1, the authors introduce some methods to improve the robustness and effectiveness of conventional fcm. based on these methods, the authors propose a novel method to solve this problem. in section 2, the authors introduce the preliminary theories of nlm filter and conventional fcm. in section 3, the authors present the motivation of non-local information and kl divergence. in section 4, the authors describe the proposed method in detail. the theoretical basis and the detailed process of the proposed method are clearly explained. in section 5, the authors introduce five evaluation indexes and show the experimental results of grayscale images and color images. in addition, the authors discuss the influence of parameter selection. in section 6, the authors point out the implications and the limitations of the method.the contributions of this work can be summarized as follows:
(1)    the proposed method uses the non-local information.
(2)    kl divergence between original membership degrees and median filtered membership degrees is considered in the method.
(3)    constraints for original fcm and non-local regularization term are adaptively specified.

the paper can be improved by the following points:
1,the authors should compare more segmentations after high-performance noise suppression methods, such as bm3d and recent deep learning based denoising methods.
2, more realistic images should be used in method evaluations.
3, computational cost and parameter sensitivity should be well analyzed
4,  some high-performance denoising methods should be analyzed in the reference:

"3d feature constrained reconstruction for low dose ct imaging," ieee transactions on circuits and systems for video technology,28(5), 1232 -1247,2018
"structure-adaptive fuzzy estimation for random-valued impulse noise suppression," ieee transactions on circuits and systems for video technology  28(2), 414 - 427,2018
"domain progressive 3d residual convolution network to improve low dose ct imaging" ieee, transactions on medical imaging, 2019.doi: 10.1109/tmi.2019.2917258
"discriminative feature representation to improve projection data inconsistency for low dose ct imaging,"ieee, transactions on medical imaging, vol. 36, no. 12, pp. 2499-2509, 2017.
. "artifact suppressed dictionary learning for low-dose ct image processing," ieee, transaction on medical imaging, 33(12), pp.2271-2292,2014

"nonlocal prior bayesian tomographic reconstruction," journal of mathematical imaging and vision, 30(2), pp.133-146, 2008.

comments from ae:

based on the reviewers' comments, i feel sorry that my recommendation is a rejection. the paper quality must be very competitive in neurocom. i do hope authors could improve their manuscript in a higher standard and welcome a new submission in the future.

回复此楼