Over the years, advancements in digital pathology have resulted in development of whole-slide imaging systems. Whole-slide images improve the quality of diagnosis by offering greater morphological details, better collaboration among the pathologist community and a unique opportunity to employ statistical and computational methods using digital image processing. However, owing to the nature of data, WSI files can be very large - anywhere between several mega-bytes to few giga-bytes – depending the magnification level and file type format. Processing such large files can be computationally demanding, especially when analysis requires higher magnification to assess histological features with better accuracy, and may become practically infeasible to analyse them on personal computers with limited resources. In this paper, we are proposing a method which will drastically reduce computation power and time consumed in WSI processing by automatically selecting the subsections of the WSI. The system produces the subsections which are more likely to encompass structures and features vital for decision making at later stage in the processing. The results section details how the qualitative analysis is unaffected while there is a drastic reduction in computation demand.
Volume 11 | Issue 9
Pages: 101-110
DOI: 10.5373/JARDCS/V11I9/20192919