Open Access

Explainable AI for binary and multi-class classification of leukemia using a modified transfer learning ensemble model

Nilkanth Mukund Deshpande

Nilkanth Mukund Deshpande

  • Department of Electronics & Telecommunication, Symbiosis Institute of Technology, Symbiosis International (Deemed University), LavalePune, India
  • Electronics & TelecommunicationAhmednagar, India
Search for this author on
Sciendo|Google Scholar
Deshpande, Nilkanth Mukund
, 
Shilpa Gite

Shilpa Gite

  • Artificial Intelligence and Machine Learning Department, Symbiosis Institute of Technology, Symbiosis International (Deemed) UniversityPune, India
  • Symbiosis Centre of Applied AI (SCAAI), Symbiosis International (Deemed) UniversityPune, India
Search for this author on
Sciendo|Google Scholar
Gite, Shilpa
 and 
Biswajeet Pradhan

Biswajeet Pradhan

  • Centre for Advanced Modelling and Geospatial Information Systems (CAMGIS), School of Civil and Environmental Engineering, Faculty of Engineering & IT, University of Technology SydneySydney, Australia
  • Earth Observation Centre, Universiti Kebangsaan, Institute of Climate ChangeBangi, Malaysia
Search for this author on
Sciendo|Google Scholar
Pradhan, Biswajeet
  
Mar 06, 2024
Explainable AI for binary and multi-class classification of leukemia using a modified transfer learning ensemble model's Cover Image
About this article
Previous Article
Next Article
Cite
Download Cover
Article Category: Article
Published Online: Mar 06, 2024
Received: Sep 07, 2023
DOI: https://doi.org/10.2478/ijssis-2024-0013
Keywords
© 2024 Nilkanth Mukund Deshpande et al., published by Sciendo
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Figure 1:

(a) Normal cells; and (b) leukemia cells [9].
(a) Normal cells; and (b) leukemia cells [9].

Figure 2:

Proposed methodology for leukemia diagnosis. LIME, Local Interpretable Model-Agnostic Explanations; XAI, explainable artificial intelligence.
Proposed methodology for leukemia diagnosis. LIME, Local Interpretable Model-Agnostic Explanations; XAI, explainable artificial intelligence.

Figure 3:

VGG-16 architecture.
VGG-16 architecture.

Figure 4:

Asymmetric convolutions.
Asymmetric convolutions.

Figure 5:

Auxiliary classifiers.
Auxiliary classifiers.

Figure 6:

Grid size reduction.
Grid size reduction.

Figure 7:

Final model architecture of inceptionv3.
Final model architecture of inceptionv3.

Figure 8:

(a) ALL-IDB-1 infected image; (b) ALL-IDB-1 normal image; (c) ALL-IDB2 infected image, and (d) ALL-IDB2 normal image.
(a) ALL-IDB-1 infected image; (b) ALL-IDB-1 normal image; (c) ALL-IDB2 infected image, and (d) ALL-IDB2 normal image.

Figure 9:

Images from the real-image dataset: (a) CLL, (b) CML, and (c) AML. AML, acute myeloid leukemia; CLL, chronic lymphocytic leukemia; CML, chronic myeloid leukemia.
Images from the real-image dataset: (a) CLL, (b) CML, and (c) AML. AML, acute myeloid leukemia; CLL, chronic lymphocytic leukemia; CML, chronic myeloid leukemia.

Figure 10:

Training and validation accuracies: (a) modified VGG-16; (b) modified Inception; and (c) ensemble model of modified InceptionNet and VGG-16 classifiers for binary classification.
Training and validation accuracies: (a) modified VGG-16; (b) modified Inception; and (c) ensemble model of modified InceptionNet and VGG-16 classifiers for binary classification.

Figure 11:

Training and validation accuracies: (a) Modified VGG-16; (b) Modified Inception; and (c) Ensemble model of modified InceptionNet and VGG-16 Classifier for multi-class classification.
Training and validation accuracies: (a) Modified VGG-16; (b) Modified Inception; and (c) Ensemble model of modified InceptionNet and VGG-16 Classifier for multi-class classification.

Figure 12:

Comparing the model accuracy with SOTA.
Comparing the model accuracy with SOTA.

Figure 13:

(a, c, e, g) Original dataset images; and (b, d, f, h) LIME interpretation results. LIME, Local Interpretable Model-Agnostic Explanations.
(a, c, e, g) Original dataset images; and (b, d, f, h) LIME interpretation results. LIME, Local Interpretable Model-Agnostic Explanations.

Metrics showing performance metric of binary and multi-class classification_

DL classifier algorithm Class/dataset Maximum training accuracy (%) Maximum validation accuracy (%)
Modified VGG-16 classifier Binary (ALL-IDB) 98.50 68.33
Modified InceptionNet classifier 98.50 78.33
Ensemble classifier 94.50 83.33
Modified VGG-16 classifier Multi-class (real-images) 98.56 93.20
Modified InceptionNet classifier 99.76 97.87
Ensemble classifier 100 100

Comparison of the proposed approach with popular SOTA_

Advantage criteria Ensemble (VGG-16 + inception) Pre-trained VGG-16 Pre-trained inception Random forest SVM ResNet50 (deep learning) EfficientNet (deep learning)
Diversity in features Yes No Yes Yes No Yes Yes
Generalization performance Good Good Good Good Good Excellent Excellent
Robustness to overfitting High High High High Moderate High High
Ensemble averaging benefit Yes No No No No No No
Feature learning capabilities Rich Deep hierarchical Diverse Moderate Linear Deep hierarchical Diverse
State-of-the-art performance Yes No (dated architecture) Yes (at the time) No No Yes Yes (as of the time of training)
Language:
English
Publication timeframe:
1 times per year
Journal Subjects:
Engineering, Introductions and Overviews, Engineering, other
Journal RSS Feed