# Artificial Intelligence Powered Automated and Early Diagnosis of Acute Lymphoblastic Leukemia Cancer in Histopathological Images: A Robust SqueezeNet-Enhanced Machine Learning Framework

## TL;DR

This paper introduces a machine learning framework using SqueezeNet to automatically detect acute lymphoblastic leukemia in histopathological images with high accuracy.

## Contribution

A novel SqueezeNet-enhanced machine learning framework for early and accurate diagnosis of acute lymphoblastic leukemia.

## Key findings

- The framework achieved 99.3% classification accuracy in detecting acute lymphoblastic leukemia.
- Validation using confusion matrix and ROC analysis confirmed the robustness of the proposed method.
- The method outperforms previous techniques in classification accuracy and reliability.

## Abstract

The growing prevalence of acute lymphoblastic leukemia cancer worldwide underlines the critical need for early and more precise detection to counter this deadly disease. This study presents a robust SqueezeNet-enhanced machine learning framework for automatically screening and classifying histopathological images for acute lymphoblastic leukemia. This work employs a deep learning (DL)–based SqueezeNet integrated with three machine learning (ML) models including neural network (NN), logistic regression (LR), and random forest (RF) for diagnosis. Combining DL and ML algorithms addresses the complexity of understanding histopathological images and the classification process. Evaluation metrics computed for acute lymphoblastic leukemia reveal a good classification accuracy (CA) of 99.3%. Results are further validated by confusion matrix (CM), calibration plot (CP), receiver operating characteristic (ROC) analysis, and comparative analysis with previous techniques. The proposed method has the potential to transform healthcare with more accurate diagnosis. It provides a robust framework for the classification of acute lymphoblastic leukemia, facilitating timely treatment options for patients.