A Review of Automatic Methods Based on Image Processing Techniques for Tuberculosis Detection from Microscopic Sputum...

Area of Science:

• Medical Microbiology
• Biomedical Engineering
• Computer Vision

Background:

• Tuberculosis (TB), caused by Mycobacterium tuberculosis, is a major global health concern, particularly in developing nations.
• Manual sputum smear microscopy, the standard TB diagnostic tool, is time-consuming and prone to errors, especially with limited trained technicians.
• The high mortality rate from TB, largely preventable with early diagnosis, necessitates improved detection methods.

Purpose of the Study:

• To review automatic methods for detecting Tuberculosis (TB) bacteria from sputum smear images.
• To analyze the evolution and accuracy of image processing techniques for TB detection between 1998 and 2014.
• To provide researchers and practitioners with an overview of TB automation advancements.

Main Methods:

• Systematic review of published automatic methods for TB detection.
• Focus on image processing techniques applied to microscopic sputum smear images.
• Analysis of studies published from 1998 to 2014.

Main Results:

• Significant improvements in the accuracy of automated TB detection algorithms over the review period.
• Emergence of commercial diagnostic products based on published research.
• Demonstrated potential of image processing for efficient and accurate TB screening.

Conclusions:

• Automated methods show considerable promise for overcoming limitations of manual TB diagnosis.
• Continued research in image processing is crucial for advancing TB detection technology.
• The development of commercial automated systems indicates the practical viability of these approaches.