Affordable Access

Access to the full text

Passive Auto Focusing of Pathological Microscope with Intelligent Field Image Collection Mechanism

Authors
  • Ghosh, Pramit1
  • Bhattacharjee, Debotosh2
  • Nasipuri, Mita2
  • 1 RCC Institute of Information Technology, Kolkata, 700015, India , Kolkata (India)
  • 2 Jadavpur University, Kolkata, 700032, India , Kolkata (India)
Type
Published Article
Journal
Journal of Medical Systems
Publisher
Springer-Verlag
Publication Date
Jan 16, 2021
Volume
45
Issue
2
Identifiers
DOI: 10.1007/s10916-020-01688-2
Source
Springer Nature
Keywords
License
Yellow

Abstract

The microscope is one of the widely used pathological equipment to analyze body fluids like blood, sputum, etc. in granular level. In order to reduce workload on pathologists and strengthen the telehealth services, an automatic self-focusing microscope with different field image collection mechanism is required. In this work, the conversion of a compound microscope into a complete digital self-focusing automatic microscope, with intelligent field image collection mechanism, is discussed. This method uses passive autofocusing technique. In this method, most informative regions are identified on the basis of texture information. Features from these identified regions are used for autofocusing the microscope. This system is capable of collecting multiple snaps from different regions of the smear sample slides. The problem with the smear slide is that it has un-uniform thickness upon the glass slide. So some region has a very thick layer and some region has a very thin layer. In general, both of these regions are not considered for pathological analysis. The proposed system is capable to detect the region of smear slide which is suitable for collection of snap images. A soft computing approach is used to detect the desired regions of the sample in the slide. The Raspberry pi is used to design the control section. Multi-threaded parallel programming is used to optimize I/O execution and waiting time. The performance of the proposed system is satisfactory. The average peak signal-to-noise ratio (PSNR) is about 33 in comparison with manual focusing by the domain expert. The performance of the system in terms of computation time, which is calculated on the benchmark microscopic image dataset, is better than other learning-based methods. Autofocusing of pathological microscope with an intelligent field image collection mechanism is highly useful in the remote healthcare domain. This work basically describes a mechanism to migrate the conventional compound microscope into a tale-health service compatible (IoT enabled) microscope. This system is highly suitable for developing countries where an overall change of existing infrastructure is difficult due to economic reasons.

Report this publication

Statistics

Seen <100 times