Exploration and Enhancement of Classifiers in the Detection of Lung Cancer from Histopathological Images
Abstract
:1. Introduction
Review of Previous Work
2. Methodology for Lung Cancer Detection
2.1. Histopathological Image Preprocessing
- The images exhibit intricate geometric structures and complex textures that arise from the vast diversity in structural morphology [37].
- Notably, histopathological images are susceptible to color inconsistencies and noise due to external factors such as variations in illumination conditions [38].
- Variations in microscope magnification, equipment settings, and other variables contribute to inconsistencies in image sizes and resolutions within histopathological images [39].
- Elements of significance, such as local micro-vessels with distinctive textural characteristics, significantly influence disease diagnosis within histopathological images. Extracting these features is of paramount importance in supporting the classification and diagnosis of lung cancer [40].
2.2. Histopathological Image Segmentation
3. Feature Extraction
3.1. Particle Swarm Optimization (PSO)
3.2. Grey Wolf Optimization (GWO)
3.3. Statistical Analysis
4. Feature Selection
4.1. KL Divergence
4.2. Invasive Weed Optimization
- Primary Population Initialization: A few seeds are dispersed to start the search.
- Reproduction process: Seeds have the potential to grow into flowering plants, which then choose and spread the fittest seeds for survival and reproduction. The quantity of grass grain grains decreases in a linear fashion from to as follows:
- Spectral Spread Method: The group’s seeds are distributed normally with a mean planting position and standard deviation (SD) determined by the equation below.
- Competitive Deprivation: If the colony has more grasses than the maximum limit (Smax), the grass with the lowest fitness is eliminated to maintain a consistent number of herbs.
- The process continues until the maximum iteration is reached, kee** the lowest cost value of the grasses.
5. Classifiers for the Detection of Lung Cancer
5.1. Support Vector Machine
5.2. K-Nearest Neighbor
5.3. Random Forest
- Randomly select M samples from × using the Bootstrap method.
- Choose n random features (where n < N) to split a decision tree node. Determine the split criterion by selecting the feature with the lowest Gini value. Gini is computed using the formula:
- Generate M decision trees by repeating steps 1 and 2, M times.
- Create a random forest by combining the decision trees and utilize voting to determine the classification outcome.
5.4. Decision Tree
5.5. Softmax Discriminant Classifier
5.6. Multilayer Perceptron
5.7. Bayesian Linear Discriminant Classifier
5.8. Methods for Updating Hyperparameters in Various Classifiers
5.8.1. Adam Approach
Algorithm 1. Adam Approach |
|
5.8.2. RAdam’s Approach
Algorithm 2. RAdam’s Approach |
|
6. Results and Discussion
6.1. Training and Testing of the Classifiers
6.2. Selection of the Optimal Parameters for the Classifiers
6.3. Performance Metrics of the Classifiers
6.4. Computational Complexity Analysis of the Classifiers
6.5. Comparison of Previous Works
7. Conclusions
Author Contributions
Funding
Institutional Review Board Statement
Informed Consent Statement
Data Availability Statement
Conflicts of Interest
References
- Prabhakar, S.K.; Lee, S.W. An Integrated Approach for Ovarian Cancer Classification with the Application of Stochastic Optimization. IEEE Access. 2020, 8, 127866–127882. [Google Scholar] [CrossRef]
- Jemal, A.; Ward, E.M.; Johnson, C.J.; Cronin, K.A.; Ma, J.; Ryerson, A.B.; Mariotto, A.; Lake, A.J.; Wilson, R.; Sherman, R.L.; et al. Annual Report to the Nation on the Status of Cancer, 1975–2014, Featuring Survival. J. Natl. Cancer Inst. 2017, 109, djx030. [Google Scholar] [CrossRef] [PubMed]
- Miki, T.; Yano, S.; Hanibuchi, M.; Sone, S. Bone Metastasis Model with Multiorgan Dissemination of Human Small-Cell Lung Cancer (SBC-5) Cells in Natural Killer Cell-Depleted SCID Mice. Oncol. Res. Featur. Preclin. Clin. Cancer Ther. 2001, 12, 209–217. [Google Scholar] [CrossRef] [PubMed]
- Mahmood, H.; Shaban, M.; Indave, B.I.; Santos-Silva, A.R.; Rajpoot, N.; Khurram, S.A. Use of artificial intelligence in diagnosis of head and neck precancerous and cancerous lesions: A systematic review. Oral Oncol. 2020, 110, 104885. [Google Scholar] [CrossRef] [PubMed]
- He, L.; Long, L.R.; Antani, S.; Thoma, G.R. Histology image analysis for carcinoma detection and grading. Comput. Methods Programs Biomed. 2012, 107, 538–556. [Google Scholar] [CrossRef] [PubMed]
- Komura, D.; Ishikawa, S. Machine Learning Methods for Histopathological Image Analysis. Comput. Struct. Biotechnol. J. 2018, 16, 34–42. [Google Scholar] [CrossRef]
- Andreadis, D.A.; Pavlou, A.M.; Panta, P. Biopsy and oral squamous cell carcinoma histopathology. In Oral Cancer Detection: Novel Strategies and Clinical Impact; Springer International Publishing: Cham, Switzerland, 2019; pp. 133–151. [Google Scholar]
- Gertych, A.; Swiderska-Chadaj, Z.; Ma, Z.; Ing, N.; Markiewicz, T.; Cierniak, S.; Salemi, H.; Guzman, S.; Walts, A.E.; Knudsen, B.S. Convolutional neural networks can accurately distinguish four histologic growth patterns of lung adenocarcinoma in digital slides. Sci. Rep. 2019, 9, 1483. [Google Scholar] [CrossRef]
- Zhang, X.; ** fire severity in south-eastern Australia using sentinel 2 and random forest. Remote Sens. Environ. 2020, 240, 111702–111715. [Google Scholar] [CrossRef]
- Zang, F.; Zhang, J. Softmax Discriminant Classifier. In Proceedings of the Third International Conference on Multimedia Information Networking and Security, Shanghai, China, 4–6 November 2011; pp. 16–19. [Google Scholar]
- Smith, K.A.; Gupta, J.N.D. Neural networks in business: Techniques and applications for the operations researcher. Comput. Oper. Res. 2000, 27, 1023–1044. [Google Scholar] [CrossRef]
- Cybenko, G.V. Approximation by superpositions of a sigmoidal function. Math. Control. Signals Syst. 1989, 2, 303–314. [Google Scholar] [CrossRef]
- Fonseca, P.; Den Teuling, N.; Long, X.; Aarts, R.M. Cardiorespiratory Sleep Stage Detection Using Conditional Random Fields. IEEE J. Biomed. Health Inform. 2017, 21, 956–966. [Google Scholar] [CrossRef]
- Claesen, M.; Moor, B.D. Hyperparameter Search in Machine Learning. ar**v 2015, ar**v:1502.02127. [Google Scholar]
- Sen, S.Y.; Ozkurt, N. Convolutional Neural Network Hyperparameter Tuning with Adam Optimizer for ECG Classification. In Proceedings of the 2020 Innovations in Intelligent Systems and Applications Conference, ASYU, Istanbul, Turkey, 15–17 October 2020; Institute of Electrical and Electronics Engineers Inc.: New York, NY, USA, 2020. [Google Scholar]
- Bukhari, S.U.K.; Syed, A.; Bokhari, S.K.A.; Hussain, S.S.; Armaghan, S.U.; Shah, S.S.H. The histological diagnosis of colonic adenocarcinoma by applying partial self supervised learning. medRxiv 2020. [Google Scholar] [CrossRef]
- Kuruvilla, J.; Gunavathi, K. Lung cancer classification using neural networks for CT images. Comput. Methods Programs Biomed. 2014, 113, 202–209. [Google Scholar] [CrossRef]
- Dabass, M.; Dabass, J. An Atrous Convolved Hybrid Seg-Net Model with residual and attention mechanism for gland detection and segmentation in histopathological images. Comput. Biol. Med. 2023, 155, 106690. [Google Scholar] [CrossRef]
- Suresh, S.; Mohan, S. ROI-based feature learning for efficient true positive prediction using convolutional neural network for lung cancer diagnosis. Neural Comput. Appl. 2020, 32, 15989–16009. [Google Scholar] [CrossRef]
- Abdul, W. An Automatic Lung Cancer Detection and Classification (ALCDC) System Using Convolutional Neural Network. In Proceedings of the 2020 13th International Conference on Developments in eSystems Engineering (DeSE), Liverpool, UK, 14–17 December 2020; pp. 443–446. [Google Scholar]
- Mastouri, R.; Khlifa, N.; Neji, H.; Hantous-Zannad, S. A bilinear convolutional neural network for lung nodules classification on CT images. Int. J. Comput. Assist. Radiol. Surg. 2021, 16, 91–101. [Google Scholar] [CrossRef] [PubMed]
- Ahmed, T.; Parvin, M.S.; Haque, M.R.; Uddin, M.S. Lung Cancer Detection Using CT Image Based on 3D Convolutional Neural Network. J. Comput. Chem. 2020, 8, 35–42. [Google Scholar] [CrossRef]
- Toğaçar, M.; Ergen, B.; Cömert, Z. Detection of lung cancer on chest CT images using minimum redundancy maximum relevance feature selection method with convolutional neural networks. Biocybern. Biomed. Eng. 2020, 40, 23–39. [Google Scholar] [CrossRef]
- Masood, A.; Sheng, B.; Li, P.; Hou, X.; Wei, X.; Qin, J.; Feng, D. Computer-Assisted Decision Support System in Pulmonary Cancer detection and stage classification on CT images. J. Biomed. Inform. 2018, 79, 117–128. [Google Scholar] [CrossRef]
- Setiawan, W.; Suhadi, M.M.; Husni; Pramudita, Y.D. Histopathology of Lung Cancer Classification Using Convolutional Neural Network with Gamma Correction. Commun. Math. Biol. Neurosci. 2022, 2022, 81. [Google Scholar]
- Pradhan, M.; Bhuiyan, A.; Mishra, S.; Thieu, T.; Coman, I.L. Histopathological Lung cancer Detection using Enhanced Grasshopper Optimization Algorithm with Random Forest. Int. J. Intell. Eng. Syst. 2022, 15, 11–20. [Google Scholar]
- Phankokkruad, M. Ensemble Transfer Learning for Lung Cancer Detection. In Proceedings of the 4th International Conferenceon Data Science and Information Technology, Shanghai, China, 23–25 July 2021; pp. 438–442. [Google Scholar]
Statistical Parameters | PSO | GWO | ||
---|---|---|---|---|
Malignant | Normal | Malignant | Normal | |
Mean | 0.8598080214 | 0.1109701363 | 0.01878313748 | 0.01751341349 |
Variance | 0.05867975074 | 0.07425036326 | 0.07492946326 | 0.07494543857 |
Skewness | 19.87029488 | 19.83047771 | 22.52231557 | 22.56212107 |
Kurtosis | 441.8828416 | 444.9961882 | 509.1565306 | 510.3537192 |
Pearson CC | 0.9019022281 | 0.9269991469 | 0.9985202125 | 0.997858273 |
CCA | 0.12309 | 0.11291 |
Classifiers | Optimal Parameters of the Classifiers |
---|---|
Support Vector Machine | Kernel—RBF; α—1; Kernel width parameter (σ)—100; w—0.85; b—0.01; Convergence Criterion—MSE. |
K-Nearest Neighbor | K—5; Distance Metric—Euclidian; w—0.5; Criterion—MSE. |
Random Forest | Number of Trees—200; Maximum Depth—10; Bootstrap Sample—20; Class Weight—0.45. |
Decision Tree | Maximum Depth—20; Impurity Criterion—MSE; Class Weight—0.4. |
Softmax Discriminant Classifier | λ = 0.5 along with mean of each class target values as 0.1 and 0.85. |
Multilayer Perceptron | Learning rate—0.3; Learning Algorithm—LM; Criterion—MSE. |
Bayesian Linear Discriminant Classifier | Prior Probability P(x)—0.5; Class mean = 0.8 and = 0.1, Criterion = MSE. |
Feature Extraction | Classifiers | Confusion Matrix | MSE | |||
---|---|---|---|---|---|---|
TP | TN | FP | FN | |||
PSO | SVM | 3944 | 4009 | 991 | 1056 | 7.29 × 10−6 |
KNN | 4267 | 3725 | 1275 | 733 | 4.49 × 10−5 | |
Random Forest | 2692 | 2933 | 2067 | 2308 | 1.60 × 10−5 | |
Decision Tree | 3184 | 3217 | 1783 | 1816 | 3.60 × 10−7 | |
Softmax Discriminant | 4033 | 3750 | 1250 | 967 | 4.00 × 10−8 | |
Multilayer Perceptron | 3425 | 3675 | 1325 | 1575 | 2.25 × 10−6 | |
Bayesian LDC | 4367 | 3975 | 1025 | 633 | 5.63 × 10−5 | |
GWO | SVM | 3617 | 4175 | 825 | 1383 | 5.76 × 10−6 |
KNN | 3500 | 3725 | 1275 | 1500 | 1.44 × 10−5 | |
Random Forest | 3967 | 3817 | 1183 | 1033 | 3.36 × 10−5 | |
Decision Tree | 4517 | 3984 | 1016 | 483 | 8.41 × 10−6 | |
Softmax Discriminant | 4083 | 4275 | 725 | 917 | 1.96 × 10−4 | |
Multilayer Perceptron | 4050 | 4384 | 616 | 950 | 4.84 × 10−4 | |
Bayesian LDC | 3967 | 3692 | 1308 | 1033 | 2.50 × 10−7 |
Feature Selection | Classifiers | Confusion Matrix | MSE | |||
---|---|---|---|---|---|---|
TP | TN | FP | FN | |||
KL Divergence | SVM | 3297 | 2747 | 2253 | 1703 | 3.24 × 10−6 |
KNN | 3978 | 2605 | 2395 | 1022 | 8.41 × 10−6 | |
Random Forest | 4115 | 3294 | 1706 | 885 | 2.30 × 10−5 | |
Decision Tree | 3919 | 4089 | 911 | 1081 | 9.00 × 10−6 | |
Softmax Discriminant | 4089 | 4258 | 742 | 911 | 4.84 × 10−6 | |
Multilayer Perceptron | 4271 | 3633 | 1367 | 729 | 2.56 × 10−6 | |
Bayesian LDC | 3298 | 3311 | 1690 | 1702 | 1.02 × 10−5 | |
IWO | SVM | 3854 | 3503 | 1497 | 1146 | 2.21 × 10−5 |
KNN | 3490 | 3985 | 1016 | 1510 | 3.36 × 10−5 | |
Random Forest | 3574 | 2757 | 2243 | 1426 | 1.94 × 10−5 | |
Decision Tree | 2982 | 2871 | 2129 | 2018 | 7.84 × 10−6 | |
Softmax Discriminant | 2734 | 3047 | 1953 | 2266 | 1.22 × 10−5 | |
Multilayer Perceptron | 3047 | 2592 | 2408 | 1953 | 1.00 × 10−6 | |
Bayesian LDC | 2681 | 2698 | 2302 | 2319 | 1.85 × 10−5 |
Feature Selection | Classifiers | Confusion Matrix | MSE | |||
---|---|---|---|---|---|---|
TP | TN | FP | FN | |||
KL Divergence | SVM | 4029 | 2742 | 2258 | 971 | 1.00 × 10−6 |
KNN | 3789 | 4147 | 853 | 1211 | 4.90 × 10−5 | |
Random Forest | 3490 | 4089 | 911 | 1510 | 6.40 × 10−7 | |
Decision Tree | 3594 | 4147 | 853 | 1406 | 2.50 × 10−7 | |
Softmax Discriminant | 4896 | 2668 | 2333 | 104 | 1.00 × 10−6 | |
Multilayer Perceptron | 3737 | 2982 | 2018 | 1263 | 2.03 × 10−5 | |
Bayesian LDC | 3460 | 2767 | 2233 | 1540 | 1.00 × 10−8 | |
IWO | SVM | 4401 | 3262 | 1738 | 599 | 4.90 × 10−7 |
KNN | 3203 | 3880 | 1120 | 1797 | 1.60 × 10−5 | |
Random Forest | 4440 | 2735 | 2265 | 560 | 1.52 × 10−5 | |
Decision Tree | 4167 | 2620 | 2380 | 833 | 5.29 × 10−6 | |
Softmax Discriminant | 4219 | 2687 | 2313 | 781 | 2.30 × 10−5 | |
Multilayer Perceptron | 4375 | 2747 | 2253 | 625 | 9.61 × 10−6 | |
Bayesian LDC | 3216 | 2760 | 2240 | 1784 | 6.89 × 10−5 |
Feature Selection | Classifiers | Confusion Matrix | MSE | |||
---|---|---|---|---|---|---|
TP | TN | FP | FN | |||
KL Divergence | SVM | 4089 | 3568 | 1433 | 911 | 6.61 × 10−4 |
KNN | 4184 | 4487 | 514 | 817 | 1.44 × 10−5 | |
Random Forest | 4555 | 3520 | 1480 | 445 | 2.72 × 10−4 | |
Decision Tree | 3815 | 3809 | 1191 | 1185 | 6.72 × 10−5 | |
Softmax Discriminant | 4392 | 3948 | 1052 | 608 | 2.40 × 10−5 | |
Multilayer Perceptron | 3881 | 4048 | 952 | 1119 | 1.96 × 10−6 | |
Bayesian LDC | 4156 | 3947 | 1053 | 844 | 8.41 × 10−6 | |
IWO | SVM | 3599 | 4085 | 915 | 1401 | 8.10 × 10−5 |
KNN | 4058 | 4375 | 625 | 942 | 7.23 × 10−5 | |
Random Forest | 4129 | 4038 | 962 | 871 | 9.00 × 10−8 | |
Decision Tree | 3713 | 4308 | 692 | 1288 | 6.40 × 10−5 | |
Softmax Discriminant | 4129 | 4161 | 839 | 871 | 4.00 × 10−4 | |
Multilayer Perceptron | 4539 | 4024 | 976 | 461 | 2.50 × 10−5 | |
Bayesian LDC | 3817 | 3797 | 1203 | 1183 | 1.44 × 10−5 |
Feature Selection | Classifiers | Confusion Matrix | MSE | |||
---|---|---|---|---|---|---|
TP | TN | FP | FN | |||
KL Divergence | SVM | 3653 | 4466 | 534 | 1347 | 1.23 × 10−5 |
KNN | 4139 | 4948 | 52 | 862 | 7.23 × 10−5 | |
Random Forest | 4044 | 3913 | 1088 | 956 | 1.30 × 10−5 | |
Decision Tree | 3635 | 3985 | 1016 | 1365 | 6.89 × 10−5 | |
Softmax Discriminant | 3565 | 4297 | 703 | 1435 | 1.37 × 10−5 | |
Multilayer Perceptron | 3740 | 4034 | 966 | 1260 | 6.40 × 10−7 | |
Bayesian LDC | 3775 | 3987 | 1013 | 1225 | 4.90 × 10−7 | |
IWO | SVM | 4339 | 4617 | 383 | 661 | 1.94 × 10−5 |
KNN | 4129 | 4321 | 680 | 871 | 5.76 × 10−6 | |
Random Forest | 4509 | 4466 | 534 | 491 | 7.57 × 10−5 | |
Decision Tree | 4617 | 4390 | 610 | 383 | 6.40 × 10−7 | |
Softmax Discriminant | 4409 | 4005 | 995 | 592 | 1.04 × 10−4 | |
Multilayer Perceptron | 4409 | 3913 | 1088 | 592 | 4.49 × 10−5 | |
Bayesian LDC | 4754 | 3973 | 1027 | 246 | 4.90 × 10−7 |
Feature Selection | Classifiers | Confusion Matrix | MSE | |||
---|---|---|---|---|---|---|
TP | TN | FP | FN | |||
KL Divergence | SVM | 4144 | 3668 | 1333 | 856 | 6.56 × 10−5 |
KNN | 4209 | 4537 | 464 | 792 | 2.92 × 10−5 | |
Random Forest | 4575 | 3620 | 1380 | 425 | 1.09 × 10−5 | |
Decision Tree | 3950 | 3859 | 1141 | 1050 | 5.93 × 10−5 | |
Softmax Discriminant | 4417 | 4098 | 902 | 583 | 1.60 × 10−5 | |
Multilayer Perceptron | 4011 | 4198 | 802 | 989 | 3.03 × 10−5 | |
Bayesian LDC | 4245 | 4047 | 953 | 755 | 3.97 × 10−5 | |
IWO | SVM | 3710 | 4235 | 765 | 1290 | 1.37 × 10−5 |
KNN | 4208 | 4375 | 625 | 792 | 4.22 × 10−5 | |
Random Forest | 4229 | 4188 | 812 | 771 | 4.49 × 10−5 | |
Decision Tree | 3813 | 4408 | 592 | 1188 | 4.36 × 10−5 | |
Softmax Discriminant | 4229 | 4211 | 789 | 771 | 1.10 × 10−4 | |
Multilayer Perceptron | 4558 | 4074 | 926 | 443 | 2.30 × 10−5 | |
Bayesian LDC | 3917 | 3897 | 1103 | 1083 | 3.02 × 10−5 |
Feature Selection | Classifiers | Confusion Matrix | MSE | |||
---|---|---|---|---|---|---|
TP | TN | FP | FN | |||
KL Divergence | SVM | 3758 | 4466 | 534 | 1242 | 1.37 × 10−5 |
KNN | 4159 | 4948 | 52 | 842 | 2.40 × 10−5 | |
Random Forest | 4094 | 4063 | 938 | 906 | 1.02 × 10−5 | |
Decision Tree | 3750 | 3985 | 1016 | 1250 | 1.23 × 10−5 | |
Softmax Discriminant | 3670 | 4297 | 703 | 1330 | 4.76 × 10−5 | |
Multilayer Perceptron | 3860 | 4084 | 916 | 1140 | 2.12 × 10−5 | |
Bayesian LDC | 3905 | 4137 | 863 | 1095 | 9.61 × 10−6 | |
IWO | SVM | 4439 | 4667 | 333 | 561 | 4.36 × 10−5 |
KNN | 4229 | 4321 | 680 | 771 | 5.48 × 10−5 | |
Random Forest | 4559 | 4466 | 534 | 441 | 1.90 × 10−4 | |
Decision Tree | 4667 | 4490 | 510 | 333 | 2.40 × 10−5 | |
Softmax Discriminant | 4459 | 4055 | 945 | 542 | 5.33 × 10−5 | |
Multilayer Perceptron | 4459 | 4063 | 938 | 542 | 5.04 × 10−5 | |
Bayesian LDC | 4789 | 4073 | 927 | 211 | 1.09 × 10−5 |
Performance Metrics | Equation | Significance |
---|---|---|
Accuracy (%) | Average positive-to-negative sample ratio. | |
Error Rate | The number of incorrect predictions, based on recorded observations. | |
F1 Score (%) | Average of precision and recall to obtain the classification accuracy of a specific class. | |
MCC | Pearson correlation between the actual output and the achieved output. | |
Jaccard Index (%) | The number of predicted true positives exceeded the number of actual positives, regardless of whether they were real or predicted. | |
g-mean (%) | Combination of sensitivity and specificity into a single value that balances both objectives. | |
Kappa | Inter-rater agreement measure for assessing agreement between two methods in categorizing cancer cases. |
Feature Extraction | Classifiers | Accuracy (%) | Error Rate (%) | F1 Score (%) | MCC | Jaccard Index (%) | g-Mean (%) | Kappa |
---|---|---|---|---|---|---|---|---|
PSO | SVM | 79.53 | 20.47 | 79.40 | 0.59 | 65.83 | 79.53 | 0.59 |
KNN | 79.92 | 20.08 | 80.95 | 0.60 | 68.00 | 80.21 | 0.60 | |
Random Forest | 56.25 | 43.75 | 55.17 | 0.13 | 38.09 | 56.26 | 0.13 | |
Decision Tree | 64.01 | 35.99 | 63.89 | 0.28 | 46.94 | 64.01 | 0.28 | |
Softmax Discriminant | 77.83 | 22.17 | 78.44 | 0.56 | 64.53 | 77.90 | 0.56 | |
Multilayer Perceptron | 71 | 29 | 70.26 | 0.42 | 54.15 | 71.04 | 0.42 | |
Bayesian LDC | 83.42 | 16.58 | 84.05 | 0.67 | 72.48 | 83.59 | 0.67 | |
GWO | SVM | 77.92 | 22.08 | 76.62 | 0.56 | 62.09 | 78.21 | 0.56 |
KNN | 72.25 | 27.75 | 71.61 | 0.45 | 55.78 | 72.29 | 0.45 | |
Random Forest | 77.84 | 22.16 | 78.17 | 0.56 | 64.16 | 77.86 | 0.56 | |
Decision Tree | 85.01 | 14.99 | 85.77 | 0.70 | 75.08 | 85.33 | 0.70 | |
Softmax Discriminant | 83.58 | 16.42 | 83.26 | 0.67 | 71.32 | 83.62 | 0.67 | |
Multilayer Perceptron | 84.34 | 15.66 | 83.80 | 0.69 | 72.12 | 84.46 | 0.69 | |
Bayesian LDC | 76.59 | 23.41 | 77.22 | 0.53 | 62.89 | 76.66 | 0.53 |
Feature Selection | Classifiers | Accuracy (%) | Error Rate (%) | F1 Score (%) | MCC | Jaccard Index (%) | g-mean (%) | Kappa |
---|---|---|---|---|---|---|---|---|
KL Divergence | SVM | 60.44 | 39.56 | 62.51 | 0.21 | 45.46 | 60.56 | 0.21 |
KNN | 65.83 | 34.17 | 69.96 | 0.33 | 53.79 | 66.96 | 0.32 | |
Random Forest | 74.09 | 25.91 | 76.05 | 0.49 | 61.36 | 74.65 | 0.48 | |
Decision Tree | 80.08 | 19.92 | 79.74 | 0.60 | 66.30 | 80.11 | 0.60 | |
Softmax Discriminant | 83.47 | 16.53 | 83.18 | 0.67 | 71.21 | 83.50 | 0.67 | |
Multilayer Perceptron | 79.04 | 20.96 | 80.30 | 0.59 | 67.08 | 79.43 | 0.58 | |
Bayesian LDC | 66.08 | 33.92 | 66.04 | 0.32 | 49.30 | 66.08 | 0.32 | |
IWO | SVM | 73.57 | 26.43 | 74.47 | 0.47 | 59.32 | 73.67 | 0.47 |
KNN | 74.74 | 25.26 | 73.43 | 0.50 | 58.01 | 74.95 | 0.49 | |
Random Forest | 63.31 | 36.69 | 66.09 | 0.27 | 49.35 | 63.64 | 0.27 | |
Decision Tree | 58.53 | 41.47 | 58.99 | 0.17 | 41.83 | 58.54 | 0.17 | |
Softmax Discriminant | 57.81 | 42.19 | 56.45 | 0.16 | 39.32 | 57.84 | 0.16 | |
Multilayer Perceptron | 56.39 | 43.61 | 58.29 | 0.13 | 41.13 | 56.44 | 0.13 | |
Bayesian LDC | 53.79 | 46.21 | 53.71 | 0.08 | 36.72 | 53.79 | 0.08 |
Feature Selection | Classifiers | Accuracy (%) | Error Rate (%) | F1 Score (%) | MCC | Jaccard Index (%) | g-mean (%) | Kappa |
---|---|---|---|---|---|---|---|---|
KL Divergence | SVM | 67.72 | 32.28 | 71.40 | 0.37 | 55.52 | 68.80 | 0.35 |
KNN | 79.36 | 20.64 | 78.60 | 0.59 | 64.74 | 79.49 | 0.59 | |
Random Forest | 75.78 | 24.22 | 74.24 | 0.52 | 59.03 | 76.09 | 0.52 | |
Decision Tree | 77.41 | 22.59 | 76.09 | 0.55 | 61.41 | 77.69 | 0.55 | |
Softmax Discriminant | 75.64 | 24.37 | 80.08 | 0.57 | 66.77 | 80.74 | 0.51 | |
Multilayer Perceptron | 67.19 | 32.81 | 69.49 | 0.35 | 53.25 | 67.54 | 0.34 | |
Bayesian LDC | 62.27 | 37.73 | 64.72 | 0.25 | 47.84 | 62.49 | 0.25 | |
IWO | SVM | 76.63 | 23.37 | 79.02 | 0.55 | 65.31 | 77.82 | 0.53 |
KNN | 70.83 | 29.17 | 68.71 | 0.42 | 52.34 | 71.16 | 0.42 | |
Random Forest | 71.75 | 28.25 | 75.87 | 0.46 | 61.12 | 74.14 | 0.43 | |
Decision Tree | 67.87 | 32.13 | 72.18 | 0.38 | 56.47 | 69.50 | 0.36 | |
Softmax Discriminant | 69.06 | 30.94 | 73.17 | 0.40 | 57.69 | 70.74 | 0.38 | |
Multilayer Perceptron | 71.22 | 28.78 | 75.25 | 0.45 | 60.32 | 73.33 | 0.42 | |
Bayesian LDC | 59.76 | 40.24 | 61.52 | 0.20 | 44.42 | 59.84 | 0.20 |
Feature Selection | Classifiers | Accuracy (%) | Error Rate (%) | F1 Score (%) | MCC | Jaccard Index (%) | g-mean (%) | Kappa |
---|---|---|---|---|---|---|---|---|
KL Divergence | SVM | 76.56 | 23.44 | 77.72 | 0.53 | 63.56 | 76.80 | 0.53 |
KNN | 86.70 | 13.30 | 86.30 | 0.74 | 75.96 | 86.81 | 0.73 | |
Random Forest | 80.75 | 19.25 | 82.56 | 0.63 | 70.30 | 81.86 | 0.62 | |
Decision Tree | 76.24 | 23.76 | 76.26 | 0.52 | 61.63 | 76.24 | 0.53 | |
Softmax Discriminant | 83.40 | 16.60 | 84.11 | 0.67 | 72.58 | 83.62 | 0.67 | |
Multilayer Perceptron | 79.28 | 20.72 | 78.93 | 0.59 | 65.20 | 79.31 | 0.59 | |
Bayesian LDC | 81.03 | 18.97 | 81.42 | 0.62 | 68.67 | 81.08 | 0.62 | |
IWO | SVM | 76.84 | 23.16 | 75.66 | 0.54 | 61.84 | 77.05 | 0.54 |
KNN | 84.33 | 15.67 | 83.82 | 0.69 | 72.14 | 84.44 | 0.69 | |
Random Forest | 81.67 | 18.33 | 81.83 | 0.63 | 69.25 | 81.68 | 0.63 | |
Decision Tree | 80.21 | 19.79 | 78.95 | 0.61 | 65.23 | 80.56 | 0.60 | |
Softmax Discriminant | 82.90 | 17.10 | 82.84 | 0.66 | 70.71 | 82.90 | 0.66 | |
Multilayer Perceptron | 85.64 | 14.36 | 86.28 | 0.72 | 75.88 | 85.94 | 0.71 | |
Bayesian LDC | 76.14 | 23.86 | 76.19 | 0.52 | 61.54 | 76.14 | 0.52 |
Feature Selection | Classifiers | Accuracy (%) | Error Rate (%) | F1 Score (%) | MCC | Jaccard Index (%) | g-mean (%) | Kappa |
---|---|---|---|---|---|---|---|---|
KL Divergence | SVM | 78.11 | 21.89 | 79.11 | 0.56 | 65.44 | 78.32 | 0.56 |
KNN | 87.45 | 12.55 | 87.02 | 0.75 | 77.03 | 87.58 | 0.75 | |
Random Forest | 81.95 | 18.05 | 83.53 | 0.65 | 71.71 | 82.92 | 0.64 | |
Decision Tree | 78.09 | 21.91 | 78.19 | 0.55 | 64.19 | 78.10 | 0.54 | |
Softmax Discriminant | 85.15 | 14.85 | 85.61 | 0.70 | 74.84 | 85.27 | 0.70 | |
Multilayer Perceptron | 82.09 | 17.91 | 81.75 | 0.64 | 69.13 | 82.12 | 0.64 | |
Bayesian LDC | 82.92 | 17.08 | 83.25 | 0.66 | 71.31 | 82.96 | 0.66 | |
IWO | SVM | 79.45 | 20.55 | 78.31 | 0.59 | 64.35 | 79.72 | 0.59 |
KNN | 85.83 | 14.17 | 85.59 | 0.72 | 74.81 | 85.86 | 0.72 | |
Random Forest | 84.17 | 15.83 | 84.23 | 0.68 | 72.76 | 84.17 | 0.68 | |
Decision Tree | 82.21 | 17.79 | 81.08 | 0.65 | 68.18 | 82.58 | 0.64 | |
Softmax Discriminant | 84.40 | 15.60 | 84.43 | 0.69 | 73.05 | 84.40 | 0.69 | |
Multilayer Perceptron | 86.32 | 13.68 | 86.95 | 0.73 | 76.91 | 86.59 | 0.73 | |
Bayesian LDC | 78.14 | 21.86 | 78.29 | 0.56 | 64.21 | 78.14 | 0.56 |
Feature Selection | Classifiers | Accuracy (%) | Error Rate (%) | F1 Score (%) | MCC | Jaccard Index (%) | g-mean (%) | Kappa |
---|---|---|---|---|---|---|---|---|
KL Divergence | SVM | 81.19 | 18.81 | 79.53 | 0.63 | 66.02 | 81.88 | 0.62 |
KNN | 90.87 | 9.14 | 90.06 | 0.83 | 81.92 | 91.71 | 0.82 | |
Random Forest | 79.56 | 20.44 | 79.83 | 0.59 | 66.43 | 79.58 | 0.59 | |
Decision Tree | 76.20 | 23.81 | 75.33 | 0.53 | 60.43 | 76.30 | 0.52 | |
Softmax Discriminant | 78.62 | 21.38 | 76.93 | 0.58 | 62.51 | 79.13 | 0.57 | |
Multilayer Perceptron | 77.74 | 22.26 | 77.07 | 0.56 | 62.69 | 77.82 | 0.55 | |
Bayesian LDC | 77.62 | 22.38 | 77.14 | 0.55 | 62.78 | 77.66 | 0.55 | |
IWO | SVM | 89.56 | 10.44 | 89.27 | 0.79 | 80.61 | 89.66 | 0.79 |
KNN | 84.50 | 15.51 | 84.19 | 0.69 | 72.70 | 84.54 | 0.69 | |
Random Forest | 89.76 | 10.24 | 89.80 | 0.80 | 81.49 | 89.76 | 0.80 | |
Decision Tree | 90.07 | 9.93 | 90.29 | 0.80 | 81.30 | 90.13 | 0.80 | |
Softmax Discriminant | 84.13 | 15.87 | 84.75 | 0.68 | 73.54 | 84.31 | 0.68 | |
Multilayer Perceptron | 83.21 | 16.79 | 84.00 | 0.67 | 72.42 | 83.47 | 0.66 | |
Bayesian LDC | 87.27 | 12.73 | 88.19 | 0.75 | 78.88 | 88.00 | 0.75 |
Feature Selection | Classifiers | Accuracy (%) | Error Rate (%) | F1 Score (%) | MCC | Jaccard Index (%) | g-mean (%) | Kappa |
---|---|---|---|---|---|---|---|---|
KL Divergence | SVM | 82.24 | 17.76 | 80.89 | 0.65 | 67.91 | 82.77 | 0.64 |
KNN | 91.07 | 8.94 | 90.30 | 0.82 | 82.31 | 91.61 | 0.82 | |
Random Forest | 81.56 | 18.44 | 81.62 | 0.63 | 68.95 | 81.56 | 0.63 | |
Decision Tree | 77.35 | 22.66 | 76.80 | 0.55 | 62.34 | 77.39 | 0.55 | |
Softmax Discriminant | 79.67 | 20.33 | 78.31 | 0.60 | 64.35 | 80.05 | 0.59 | |
Multilayer Perceptron | 79.44 | 20.56 | 78.97 | 0.59 | 65.25 | 79.49 | 0.59 | |
Bayesian LDC | 80.42 | 19.58 | 79.96 | 0.61 | 66.61 | 80.47 | 0.61 | |
IWO | SVM | 91.06 | 8.94 | 90.86 | 0.82 | 83.24 | 91.13 | 0.82 |
KNN | 85.50 | 14.51 | 85.36 | 0.71 | 74.46 | 85.50 | 0.71 | |
Random Forest | 90.26 | 9.74 | 90.35 | 0.81 | 82.39 | 90.27 | 0.81 | |
Decision Tree | 91.57 | 8.43 | 91.71 | 0.83 | 84.70 | 91.87 | 0.83 | |
Softmax Discriminant | 85.13 | 14.87 | 85.71 | 0.70 | 75.00 | 85.32 | 0.70 | |
Multilayer Perceptron | 85.21 | 14.79 | 85.77 | 0.71 | 75.09 | 85.39 | 0.70 | |
Bayesian LDC | 88.62 | 11.38 | 89.38 | 0.78 | 80.80 | 89.25 | 0.77 |
S No | Feature Extraction | Feature Selection | Classifiers | Accuracy (%) |
---|---|---|---|---|
1 | PSO | - | Bayesian LDC | 83.42% |
2 | GWO | - | Decision Tree | 85.01% |
3 | PSO | KL Divergence | Softmax Discriminant | 83.47% |
4 | PSO | IWO | KNN | 74.74% |
5 | GWO | KL Divergence | KNN | 79.36% |
6 | GWO | IWO | SVM | 76.63% |
7 | PSO | KL Divergence | KNN with Adam | 86.70% |
8 | PSO | IWO | MLP with Adam | 85.64% |
9 | PSO | KL Divergence | KNN with RAdam | 87.45% |
10 | PSO | IWO | MLP with RAdam | 86.32% |
11 | GWO | KL Divergence | KNN with Adam | 90.87% |
12 | GWO | IWO | Decision Tree with Adam | 90.07% |
13 | GWO | KL Divergence | KNN with RAdam | 91.07% |
14 | GWO | IWO | Decision Tree with RAdam | 91.57% |
S No | Classifiers | Without Feature Extraction | With Feature Extraction | With Feature Selection | With Hyperparameter Tuning of IWO Feature Selection Method | |||
---|---|---|---|---|---|---|---|---|
PSO | GWO | KL Divergence | IWO | Adam | RAdam | |||
1 | SVM | |||||||
2 | KNN | |||||||
3 | RF | |||||||
4 | DT | |||||||
5 | SDC | |||||||
6 | MLP | |||||||
7 | BLDC |
S No | Authors | Dataset Used | Machine Learning Models/Classifiers | Accuracy (%) |
---|---|---|---|---|
1 | Bukhari, S. et al. [57] | CRAG Dataset | ResNet-50 | 93.91% |
2 | **sa Kuruvilla. et al. [58] | LIDC Dataset (155 Patients—CT images) | Feed Forward Back Propagation Neural Networks | 93.3% |
3 | Dabass, M. et al. [59] | CRAG Dataset | Atrous Convolved Hybrid Seg-Net Architecture | 87.63% |
4 | Supriya, Suresh.et al. [60] | LIDC-IDRI Repository (CT scans) | CNN | 93.9% |
5 | Wadood, Abdul. [61] | LIDC-IDRI Repository (CT scans) | CNN-ALCDC | 97.2% |
6 | Rekka, Mastouri. et al. [62] | LUNA16 Database (3186 CT images) | BCNN [VGG16, VGG19] | 91.99% |
7 | Tasnim, Ahmed. et al. [63] | LUNA16 Database | 3D CNN | 80% |
8 | Mesut Toğaçar. et al. [64] | Cancer Imaging Archieve (CT images) | AlexNet and kNN classifier | 98.74% |
9 | Anum, Masood. et al. [65] | Biomedical Datasets–IoT | CNN DFCNet | 77.6% 84.58% |
10 | Wahyudi, Setiawan. et al. [66] | LC25000 Database | CNN | 87.16% |
11 | Manaswini, Pradhan. [67] | LC25000 Database | Without Feature Selection (EGOA)—KNN With Feature Selection (EGOA)–KNN | 80.16% 81.59% |
12 | Phankokkruad, M [68] | LC25000 Database | Ensemble ResNet50V2 | 91% 90% |
13 | Karthikeyan Shanmugam, Harikumar Rajaguru | LC25000 Database | Feature Extraction-GWO Feature Selection—IWO Decision tree with RAdam Hyper parameter Updation method | 91.57% |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Shanmugam, K.; Rajaguru, H. Exploration and Enhancement of Classifiers in the Detection of Lung Cancer from Histopathological Images. Diagnostics 2023, 13, 3289. https://doi.org/10.3390/diagnostics13203289
Shanmugam K, Rajaguru H. Exploration and Enhancement of Classifiers in the Detection of Lung Cancer from Histopathological Images. Diagnostics. 2023; 13(20):3289. https://doi.org/10.3390/diagnostics13203289
Chicago/Turabian StyleShanmugam, Karthikeyan, and Harikumar Rajaguru. 2023. "Exploration and Enhancement of Classifiers in the Detection of Lung Cancer from Histopathological Images" Diagnostics 13, no. 20: 3289. https://doi.org/10.3390/diagnostics13203289