Lung and Colon Cancer Classification of Histopathology Images using ImageNet - Pretrained EfficientNetB4 with MLP Head

DOI

10.2991/978-94-6239-616-6_52How to use a DOI?

Keywords

Lung cancer; Colon cancer; Pretrained – ImageNet; EfficientNetB4; MLP classifier head

Abstract

Lung and colon cancers are among the leading cancers worldwide and are major contributors to mortality since they are often diagnosed at a late stage. Early and precise histopathological classification is essential for determining appropriate treatment options. In this work, we propose a lightweight yet effective framework that integrates the ImageNet-pretrained EfficientNet-B4 backbone with a compact Multilayer Perceptron (MLP) classifier head, offering a simpler and computationally efficient alternative to existing deep or ensemble models. Unlike previous studies that rely on multiple pretrained networks or complex ensembles, the proposed approach maintains high accuracy with significantly reduced model and training cost. To evaluate the proposed work, the model was trained to classify histopathological images into five distinct groups: lung benign, lung adenocarcinoma, lung squamous cell carcinoma, colon adenocarcinoma, and colon benign. A total of 25,000 images were used, including data preprocessing and augmentation steps. The final model achieved a testing accuracy of 96%, demonstrating per-class accuracies of 0.98 for colon adenocarcinoma, 1.00 for colon benign, 1.00 for lung benign, 0.88 for lung adenocarcinoma and 0.95 for lung squamous cell carcinoma. Precision and recall both reached 1.00 for colon and lung benign tissues, while the F1-score for lung adenocarcinoma and lung squamous cell carcinoma was 0.92. These results indicate that the EfficientNet-B4 + MLP architecture achieves a reasonable balance between high accuracy and computational efficiency and thus offers a valuable baseline for clinical histopathological applications.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - R. Yogalakshmi
AU  - S. Shri Vatssan
AU  - S. Bala Abinaya
AU  - R. Sathishkumar
PY  - 2026
DA  - 2026/03/31
TI  - Lung and Colon Cancer Classification of Histopathology Images using ImageNet - Pretrained EfficientNetB4 with MLP Head
BT  - Proceedings of the International Conference on Artificial Intelligence and Secure Data Analytics (ICAISDA 2025)
PB  - Atlantis Press
SP  - 695
EP  - 704
SN  - 1951-6851
UR  - https://doi.org/10.2991/978-94-6239-616-6_52
DO  - 10.2991/978-94-6239-616-6_52
ID  - Yogalakshmi2026
ER  -