pith. sign in

arxiv: 1907.09505 · v1 · pith:LKC24XHBnew · submitted 2019-07-22 · 📡 eess.IV

Improving Brain Magnetic Resonance Image MRI Segmentation via a Novel Algorithm based on Genetic and Regional Growth

Amir Javadpour , Alireza Mohammadi This is my paper

Pith reviewed 2026-05-24 17:33 UTC · model grok-4.3

classification 📡 eess.IV
keywords brain MRI segmentationgenetic algorithmregional growthmedical image processingAlzheimer's diagnosisimage segmentationautomatic parameter selection
0
0 comments X

The pith

Genetic algorithm automates initial point selection for regional growth to segment brain MRIs with lower error.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces a segmentation approach for brain MRI scans that pairs regional growth with a genetic algorithm to automatically pick initial pixels and similarity criteria. This automation targets higher accuracy than manual parameter choices in analyzing tissue size variations tied to conditions like Alzheimer's. A sympathetic reader would care because reliable segmentation serves as an early step in disease diagnosis from non-invasive images. The authors apply the method to sample images and report reduced segmentation errors relative to standard regional growth with manual starts.

Core claim

By using genetic algorithms to automatically select primary pixels and similarity criteria for the regional growth method, the proposed algorithm improves MRI segmentation accuracy and reduces error compared to manual selection of initial points.

What carries the argument

Genetic algorithm that optimizes initial points and similarity criteria to drive the regional growth segmentation process.

If this is right

  • Segmentation error decreases when initial points come from genetic optimization rather than manual choice.
  • The method supports more consistent identification of brain tissue changes linked to diseases.
  • Automatic selection of similarity criteria improves the validity of the resulting segmentations.
  • The combined approach applies directly to existing brain MRI datasets without extra manual setup.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same automation of starting parameters could extend to segmenting other soft-tissue medical scans beyond brain MRIs.
  • Comparison against contemporary machine-learning segmentation tools on the same images would test relative performance.
  • Wider use might lower reliance on expert radiologists for initial parameter setting in routine scans.

Load-bearing premise

The genetic algorithm's automatic choices of initial points and similarity criteria produce objectively better segmentation than manual selection without needing further tuning for the images tested.

What would settle it

A head-to-head test on new brain MRI images where the genetic algorithm version shows equal or higher segmentation error than conventional regional growth with manual initial points.

Figures

Figures reproduced from arXiv: 1907.09505 by Alireza Mohammadi, Amir Javadpour.

Figure 1
Figure 1. Figure 1: Digital Image of Brain 97 [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: MRI Segmentation into Gray Matter, White Matter and Cerebrospinal Fluid Compo￾nents Using Region Growth Method 99 [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Applying k-means Algorithms on Brain Image (A): Original Image, (B): Result￾ed Image Advantages Regions having desired specifications are correctly separated. Good segmentation results for image have distinctive edge. It is simple and develops only by some initial grains. One can select initial grains and desired criterion. Several criteria can be used simultaneously. There is good performance with respect… view at source ↗
Figure 4
Figure 4. Figure 4: Steps of Growing Regional Algo￾rithm 101 [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Four-neighbor Growth [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Diagonal Four-neighbor Growth [PITH_FULL_IMAGE:figures/full_fig_p008_6.png] view at source ↗
Figure 8
Figure 8. Figure 8: shows one sample image from this database. Gray matter, white matter and cere￾brospinal fluid images pertaining to [PITH_FULL_IMAGE:figures/full_fig_p008_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Individual Image of Gray Matter, White Matter and Cerebrospinal Fluid pertaining to [PITH_FULL_IMAGE:figures/full_fig_p009_9.png] view at source ↗
Figure 10
Figure 10. Figure 10 [PITH_FULL_IMAGE:figures/full_fig_p009_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: shows the result of applying pro￾posed algorithm on one of the images in this database. As can be seen, the proposed method has segmented the image into 3 components. Results for comparison of proposed method and original images of database are listed in [PITH_FULL_IMAGE:figures/full_fig_p010_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Results from Application of Pro￾posed Algorithms on a Typical Image taken from BrainWeb Database Improving Brain Magnetic Resonance Image Segmentation 105 [PITH_FULL_IMAGE:figures/full_fig_p011_12.png] view at source ↗
read the original abstract

Background: Regarding the importance of right diagnosis in medical applications, various methods have been exploited for processing medical images solar. The method of segmentation is used to analyze anal to miscall structures in medical imaging. Objective: This study describes a new method for brain Magnetic Resonance Image (MRI) segmentation via a novel algorithm based on genetic and regional growth. Methods: Among medical imaging methods, brains MRI segmentation is important due to the high contrast of non-intrusive soft tissue and high spatial resolution. Size variations of brain tissues are often accompanied by various diseases such as Alzheimers disease. As our knowledge about the relationship between various brain diseases and deviation of brain anatomy increases, MRI segmentation is exploited as the first step in early diagnosis. In this paper, the regional growth method and auto-mate selection of initial points by genetic algorithm are used to introduce a new method for MRI segmentation. Primary pixels and similarity criterion are automatically by genetic algorithms to maximize the accuracy and validity in image segmentation. Results: By using genetic algorithms and defining the fixed function of image segmentation, the initial points for the algorithm were found. The proposed algorithms are applied to the images and results are manually selected by regional growth in which the initial points were compared. The results showed that the proposed algorithm could reduce segmentation error effectively. Conclusion: The study concluded that the proposed algorithm could reduce segmentation error effectively and help us to diagnose brain diseases.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript proposes a hybrid segmentation method for brain MRI that uses a genetic algorithm to automatically select initial seed points and similarity criteria for a subsequent region-growing procedure. The central claim is that this automation yields measurably lower segmentation error than conventional region growing and thereby assists diagnosis of brain diseases such as Alzheimer’s.

Significance. If the claimed error reduction were demonstrated with quantitative, reproducible metrics and independent validation, the approach could supply a practical, parameter-light alternative for automated MRI analysis. No such demonstration is present, so the potential clinical utility cannot be assessed.

major comments (2)
  1. [Results] Results paragraph: the assertion that 'the results showed that the proposed algorithm could reduce segmentation error effectively' is unsupported by any numerical evidence. No Dice/Jaccard scores, pixel-wise error rates, subject counts, baseline comparisons, or statistical tests appear anywhere in the manuscript.
  2. [Results] Results paragraph: the evaluation is described as 'results are manually selected by regional growth in which the initial points were compared,' indicating only subjective visual inspection. This directly undermines the central claim of objective superiority.
minor comments (2)
  1. [Abstract] Abstract contains multiple typographical and grammatical errors ('medical images solar', 'analyze anal to miscall structures', 'auto-mate selection') that must be corrected for readability.
  2. [Methods] No description is given of the genetic-algorithm fitness function, population size, selection/crossover/mutation operators, or termination criteria, rendering the method non-reproducible.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for identifying key deficiencies in the quantitative validation of our proposed method. We agree that the current manuscript lacks the necessary numerical evidence and objective metrics to support its claims, and we will substantially revise the results section in the next version.

read point-by-point responses

  1. Referee: [Results] Results paragraph: the assertion that 'the results showed that the proposed algorithm could reduce segmentation error effectively' is unsupported by any numerical evidence. No Dice/Jaccard scores, pixel-wise error rates, subject counts, baseline comparisons, or statistical tests appear anywhere in the manuscript.

    Authors: We agree that the manuscript provides no quantitative support for the claim. The revised version will include a full results section reporting Dice coefficients, Jaccard indices, pixel-wise error rates, the number of subjects/images used, direct comparisons against standard region-growing baselines, and statistical significance tests. revision: yes

  2. Referee: [Results] Results paragraph: the evaluation is described as 'results are manually selected by regional growth in which the initial points were compared,' indicating only subjective visual inspection. This directly undermines the central claim of objective superiority.

    Authors: The manuscript text does describe a manual, subjective comparison process. This is a valid and important criticism. In revision we will replace all subjective visual assessments with the quantitative, reproducible metrics and baseline comparisons noted above. revision: yes

Circularity Check

0 steps flagged

No circularity in derivation chain

full rationale

The paper proposes an algorithmic method combining genetic algorithms for automatic seed-point and similarity-criterion selection with region-growing segmentation. Its central claim is an empirical assertion that the approach reduces segmentation error, supported only by a qualitative statement that results were 'manually selected' after application to images. No equations, first-principles derivations, or mathematical steps are presented whose outputs reduce to the inputs by construction. No self-citations, uniqueness theorems, or ansatzes are invoked. The absence of quantitative metrics or baselines is a limitation of evidence, not a circular reduction of any claimed derivation to its own fitted values. The paper therefore contains no load-bearing circular steps.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract supplies insufficient technical detail to enumerate specific free parameters, axioms, or invented entities; the approach rests on the standard assumptions of genetic algorithms and region growing without additional specification.

pith-pipeline@v0.9.0 · 5784 in / 957 out tokens · 29219 ms · 2026-05-24T17:33:16.521410+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

37 extracted references · 37 canonical work pages

  1. [1]

    Adaptive Markov modeling for mutual-informa- tion-based, unsupervised MRI brain-tissue clas- sification

    Awate SP, Tasdizen T, Foster N, Whitaker RT. Adaptive Markov modeling for mutual-informa- tion-based, unsupervised MRI brain-tissue clas- sification. Med Image Anal. 2006;10:726-39. doi. org/10.1016/j.media.2006.07.002. PubMed PMID: 16919993

    work page doi:10.1016/j.media.2006.07.002 2006

  2. [2]

    Implementing a Smart Method to Eliminate Artifacts of Vital Sig- nals

    Javadpour A, Mohammadi A. Implementing a Smart Method to Eliminate Artifacts of Vital Sig- nals. J Biomed Phys Eng. 2015; 5(4): 199–206. PMCID: PMC4681465

    work page 2015

  3. [3]

    Diagnosis of Al- zheimer’s disease from EEG signals: where are we standing? Curr Alzheimer Res

    Dauwels J, Vialatte F, Cichocki A. Diagnosis of Al- zheimer’s disease from EEG signals: where are we standing? Curr Alzheimer Res. 2010;7:487-505. doi.org/10.2174/156720510792231720. PubMed PMID: 20455865

    work page doi:10.2174/156720510792231720 2010

  4. [4]

    Computation in neurons and neural systems: Springer Science & Business Media; 2012

    Eeckman F. Computation in neurons and neural systems: Springer Science & Business Media; 2012

    work page 2012

  5. [5]

    Automated Region Growing for Segmentation of Brain Lesion in Diffusion- weighted MRI

    Saad NM, Abu-Bakar S, Muda S, Mokji M, Abdullah A, editors. Automated Region Growing for Segmentation of Brain Lesion in Diffusion- weighted MRI. Proceedings of the International MultiConference of Engineers and Computer Scientists; 2012

    work page 2012

  6. [6]

    Masset, R

    Banerjee S. The macroeconomics of dementia- -will the world economy get Alzheimer’s disease? Arch Med Res. 2012;43:705-9. doi.org/10.1016/j. arcmed.2012.10.006. PubMed PMID: 23085453

    work page doi:10.1016/j 2012

  7. [7]

    Epidemiology of Alzheimer’s Disease: INTECH Open Access Pub- lisher; 2013

    Xu W, Ferrari C, Wang HX. Epidemiology of Alzheimer’s Disease: INTECH Open Access Pub- lisher; 2013. Javadpour A., Mohammadi A. 106 J Biomed Phys Eng 2016; 6(2) www.jbpe.org

    work page 2013

  8. [8]

    MRI tissue classification and bias field estimation based on coherent local intensity clustering: A unified energy minimization framework

    Li C, Xu C, Anderson AW, Gore JC, editors. MRI tissue classification and bias field estimation based on coherent local intensity clustering: A unified energy minimization framework. Informa- tion Processing in Medical Imaging: Springer

    work page

  9. [9]

    Segmentation of Brain MRI: INTECH Open Access Publisher; 2012

    Xu R, Ohya J, Luo L. Segmentation of Brain MRI: INTECH Open Access Publisher; 2012

    work page 2012

  10. [10]

    Neuroanatomical segmentation in MRI: technological objectives

    Worth AJ, Makris N, Caviness Jr VS, Kennedy DN. Neuroanatomical segmentation in MRI: technological objectives. International Journal of Pattern Recognition and Artificial Intel- ligence. 1997;11:1161-87. doi.org/10.1142/ S0218001497000548

    work page 1997

  11. [11]

    Integrated graph cuts for brain MRI segmentation

    Song Z, Tustison N, Avants B, Gee JC. Integrated graph cuts for brain MRI segmentation. Medical Image Computing and Computer-Assisted Inter- vention–MICCAI 2006: Springer; 2006. p. 831-8

    work page 2006

  12. [12]

    MRI brain volume abnormalities in young, nonpsychotic relatives of schizophrenia pro- bands are associated with subsequent prodromal symptoms

    Ho BC. MRI brain volume abnormalities in young, nonpsychotic relatives of schizophrenia pro- bands are associated with subsequent prodromal symptoms. Schizophr Res. 2007;96:1-13. doi. org/10.1016/j.schres.2007.08.001. PubMed PMID: 17761401. PubMed PMCID: 2222920

    work page doi:10.1016/j.schres.2007.08.001 2007

  13. [13]

    Slezak D, Pal S, Kang BH, Gu J, Kuroda H, Kim TH. Signal Processing, Image Processing and Pattern Recognition: International Conference, SIP 2009, Held as Part of the Future Generation Infor- mation Technology Conference, FGIT 2009, Jeju Island, Korea, December 10-12, 2009. Proceed- ings: Springer; 2010

    work page 2009

  14. [14]

    Segmentation of white matter lesions from volumetric MR images

    Hojjatoleslami S, Kruggel F, von Cramon D, edi- tors. Segmentation of white matter lesions from volumetric MR images. Medical Image Computing and Computer-Assisted Intervention–MICCAI’99: Springer; 1999. p. 52-61

    work page 1999

  15. [15]

    Magnetic resonance imaging (MRI): considerations and applications in radiotherapy treatment planning

    Khoo VS, Dearnaley DP, Finnigan DJ, Padhani A, Tanner SF, Leach MO. Magnetic resonance imaging (MRI): considerations and applications in radiotherapy treatment planning. Radiother Oncol. 1997;42:1-15. doi.org/10.1016/S0167- 8140(96)01866-X. PubMed PMID: 9132820

    work page doi:10.1016/s0167- 1997

  16. [16]

    A Fuzzy Relative of the ISODATA Pro- cess and Its Use in Detecting Compact Well-Sepa- rated Clusters

    Dunn JC. A Fuzzy Relative of the ISODATA Pro- cess and Its Use in Detecting Compact Well-Sepa- rated Clusters. Journal of Cybernetics. 1973;3:32-

    work page 1973

  17. [17]

    doi.org/10.1080/01969727308546046

    work page doi:10.1080/01969727308546046

  18. [18]

    Pattern recognition with fuzzy objec- tive function algorithms: Springer Science & Business Media; 2013

    Bezdek JC. Pattern recognition with fuzzy objec- tive function algorithms: Springer Science & Business Media; 2013

    work page 2013

  19. [19]

    The possibilistic c-means algorithm: insights and recommenda- tions

    Krishnapuram R, Keller JM. The possibilistic c-means algorithm: insights and recommenda- tions. Fuzzy Systems, IEEE Transactions on. 1996;4:385-93. doi.org/10.1109/91.531779

    work page doi:10.1109/91.531779 1996

  20. [20]

    Improved possibilistic c-means clustering algorithms

    Zhang JS, Leung YW. Improved possibilistic c-means clustering algorithms. Fuzzy Systems, IEEE Transactions on. 2004;12:209-17. doi. org/10.1109/TFUZZ.2004.825079

    work page doi:10.1109/tfuzz.2004.825079 2004

  21. [21]

    An improved MR image segmentation method based on fuzzy c- means clustering

    Lu L, Li M, Zhang X, editors. An improved MR image segmentation method based on fuzzy c- means clustering. Computational Problem-Solving (ICCP), 2012 International Conference on; 2012: IEEE

    work page 2012

  22. [22]

    A fully automated algorithm under modified FCM frame- work for improved brain MR image segmentation

    Sikka K, Sinha N, Singh PK, Mishra AK. A fully automated algorithm under modified FCM frame- work for improved brain MR image segmentation. Magn Reson Imaging. 2009;27:994-1004. doi. org/10.1016/j.mri.2009.01.024. PubMed PMID: 19395212

    work page doi:10.1016/j.mri.2009.01.024 2009

  23. [23]

    Seed-based region growing study for brain abnormalities segmentation

    Khalid NEA, Ibrahim S, Manaf M, Ngah UK, edi- tors. Seed-based region growing study for brain abnormalities segmentation. Information Technol- ogy (ITSim), 2010 International Symposium in; 2010: IEEE. doi.org/10.1109/itsim.2010.5561560

    work page doi:10.1109/itsim.2010.5561560 2010

  24. [24]

    Unifying variational approach and region growing segmentation

    Rose J, Grenier T, Revol-Muller C, Odet C, edi- tors. Unifying variational approach and region growing segmentation. European Signal Process- ing Conference; 2010

    work page 2010

  25. [25]

    Genetic based Fuzzy Seeded Region Growing Segmenta- tion for diabetic retinopathy images

    Tamilarasi M, Duraiswamy K, editors. Genetic based Fuzzy Seeded Region Growing Segmenta- tion for diabetic retinopathy images. Computer Communication and Informatics (ICCCI), 2013 International Conference on; 2013: IEEE. doi. org/10.1109/iccci.2013.6466117

    work page doi:10.1109/iccci.2013.6466117 2013

  26. [26]

    Brain tumor segmentation from MRI data sets using region growing approach

    Węgliński T, Fabijańska A, editors. Brain tumor segmentation from MRI data sets using region growing approach. Perspective Technologies and Methods in MEMS Design (MEMSTECH), 2011 Proceedings of VIIth International Conference on; 2011: IEEE

    work page 2011

  27. [27]

    Survey: interpolation methods in medical image process- ing

    Lehmann TM, Gonner C, Spitzer K. Survey: interpolation methods in medical image process- ing. IEEE Trans Med Imaging. 1999;18:1049-75. doi.org/10.1109/42.816070. PubMed PMID: 10661324

    work page doi:10.1109/42.816070 1999

  28. [28]

    A survey of thresholding techniques

    Sahoo PK, Soltani S, Wong AK. A survey of thresholding techniques. Computer vision, graph- ics, and image processing. 1988;41:233-60. doi. org/10.1016/0734-189X(88)90022-9

    work page doi:10.1016/0734-189x(88)90022-9 1988

  29. [29]

    Neural network for enhancement of FCM based brain MRI segmentation

    Rostami MT, Ghasemi J, Ghaderi R, editors. Neural network for enhancement of FCM based brain MRI segmentation. Fuzzy Systems (IFSC), 2013 13th Iranian Conference on; 2013: IEEE. doi. org/10.1109/ifsc.2013.6675661

    work page doi:10.1109/ifsc.2013.6675661 2013

  30. [30]

    Watersheds in digital spaces: an efficient algorithm based on immersion simu- lations

    Vincent L, Soille P. Watersheds in digital spaces: an efficient algorithm based on immersion simu- lations. IEEE Transactions on Pattern Analysis & Machine Intelligence. 1991;(6):583-98. doi. org/10.1109/34.87344

    work page doi:10.1109/34.87344 1991

  31. [31]

    Trémeau A, Tominaga S, Plataniotis KN. Color in image and video processing: most recent Improving Brain Magnetic Resonance Image Segmentation 107 J Biomed Phys Eng 2016; 6(2) www.jbpe.org trends and future research directions. Journal on Image and Video Processing. 2008;2008:7. doi. org/10.1155/2008/581371

    work page doi:10.1155/2008/581371 2016

  32. [32]

    Image seg- mentation using automatic seeded region growing and instance-based learning

    Gómez O, González JA, Morales EF. Image seg- mentation using automatic seeded region growing and instance-based learning. Progress in pattern recognition, image analysis and applications: Springer; 2007. p. 192-201

    work page 2007

  33. [33]

    Available from: http//brainweb.bic.mni

    BrainWeb: 20 Anatomical Models of 20 Normal Brains. Available from: http//brainweb.bic.mni. mcgill.ca/brainweb/anatomic_normal_20

    work page

  34. [34]

    Real-time ocular artifact suppression using recurrent neural network for electro-encephalogram based brain-computer in- terface

    Erfanian A, Mahmoudi B. Real-time ocular artifact suppression using recurrent neural network for electro-encephalogram based brain-computer in- terface. Med Biol Eng Comput. 2005;43:296-305. doi.org/10.1007/BF02345969. PubMed PMID: 15865142

    work page doi:10.1007/bf02345969 2005

  35. [35]

    Automatic Sleep Stages Detection Based on EEG Signals Using Combination of Classifiers

    Kianzad R, Montazery Kordy H. Automatic Sleep Stages Detection Based on EEG Signals Using Combination of Classifiers. Journal of Electri- cal and Computer Engineering Innovations. 2013;1:99-105

    work page 2013

  36. [36]

    A New Blind Source Separation Method to Remove Artifact in EEG Signals

    Chaozhu Z, Siyao L, Abdullah AK, editors. A New Blind Source Separation Method to Remove Artifact in EEG Signals. Instrumentation, Measure- ment, Computer, Communication and Control (IMCCC), 2013 Third International Conference on; 2013: IEEE. doi.org/10.1109/imccc.2013.319

    work page doi:10.1109/imccc.2013.319 2013

  37. [37]

    Detection and extraction of the ECG signal parameters

    Gholam-Hosseini H, Nazeran H, editors. Detection and extraction of the ECG signal parameters. En- gineering in Medicine and Biology Society, 1998. Proceedings of the 20th Annual International Con- ference of the IEEE; 1998: IEEE. doi.org/10.1109/ iembs.1998.745846. Javadpour A., Mohammadi A. 108

    work page arXiv 1998