# White blood cell subtype detection and classification

**Authors:** Nalla Praveen, Narinder Singh Punn, Sanjay Kumar Sonbhadra, Sonali Agarwal, M. Syafrullah, Krisna Adiyarta  
**arXiv:** [2108.04614](https://arxiv.org/abs/2108.04614)  
**Date:** 2021-08-10  
**Summary page:** https://www.summarizepaper.com/en/paper/2108.04614v3/

## Key points

- Machine learning has had a significant impact on the healthcare industry, particularly in transforming medical diagnosis and treatment.
- White blood cell classification is a crucial aspect of accurate medical diagnosis traditionally done manually by haematologists, but it is time-consuming and prone to errors.
- Researchers are now using computer-aided detection and deep learning approaches to automate white blood cell classification processes.
- A recent study proposed using YOLOv3 object detection technique to both localize and classify white blood cells with bounding boxes, achieving 99.2% accuracy in detecting cells and 90% accuracy in classifying them.
- This advancement represents a significant step forward in automating and improving the efficiency of white blood cell classification processes, contributing valuable knowledge to healthcare diagnostics.

## Summary

, , , , 

Machine learning has greatly impacted the healthcare industry, offering numerous possibilities for transforming medical diagnosis and treatment. One promising area of research is white blood cell classification, which plays a vital role in accurate medical diagnosis. Traditionally, this task is performed by haematologists through manual examination of a small smear of blood under a microscope. However, this process is time-consuming and prone to human error. To overcome these challenges, researchers have turned to computer-aided detection and diagnosis using deep learning approaches. While these methods have shown success in identifying white blood cell subtypes, they often struggle with accurately localizing the positions of cells within images. In response to this limitation, a recent study proposed utilizing the YOLOv3 object detection technique to both localize and classify white blood cells with bounding boxes. The study conducted exhaustive experimental analysis and achieved an impressive 99.2% accuracy in detecting white blood cells and 90% accuracy in classifying them. This advancement represents a significant step forward in automating and improving the efficiency of white blood cell classification processes. In addition to showcasing their methodology's technical aspects, the paper also provides insights into related work in the field of research and outlines future directions for enhancing white blood cell subtype detection and classification. Overall, this research contributes valuable knowledge to ongoing efforts to leverage machine learning for advancements in healthcare diagnostics.