Ensembling Handcrafted Features with Deep Features: An Analytical Study for Classification of Routine Colon Cancer Histopathological Nuclei Images

TL;DR

This study investigates combining handcrafted features with deep learning models to improve classification of colon cancer histopathological images, highlighting the benefits and limitations of ensemble approaches in complex medical datasets.

Contribution

The paper provides an analytical comparison of ensemble methods combining handcrafted and deep features, emphasizing the importance of dataset-specific analysis for optimal model performance.

Findings

Ensembling deep features improves classification accuracy.

Handcrafted features alone show limited impact on deep model performance.

Proper dataset analysis is crucial for effective model deployment.

Abstract

The use of Deep Learning (DL) based methods in medical histopathology images have been one of the most sought after solutions to classify, segment, and detect diseased biopsy samples. However, given the complex nature of medical datasets due to the presence of intra-class variability and heterogeneity, the use of complex DL models might not give the optimal performance up to the level which is suitable for assisting pathologists. Therefore, ensemble DL methods with the scope of including domain agnostic handcrafted Features (HC-F) inspired this work. We have, through experiments, tried to highlight that a single DL network (domain-specific or state of the art pre-trained models) cannot be directly used as the base model without proper analysis with the relevant dataset. We have used F1-measure, Precision, Recall, AUC, and Cross-Entropy Loss to analyse the performance of our approaches. We observed from the results that the DL features ensemble bring a marked improvement in the overall performance of the model, whereas, domain agnostic HC-F remains dormant on the performance of the DL models.