QUANTIFICATION OF TYPE III COLLAGEN DEPOSITION DENSITY FROM PHOTOMICROGRAPH OF VAGINAL CONNECTIVE TISSUE

• Main Author: Arfan, Muhammad
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/66640
• Institution: Institut Teknologi Bandung

ABSTRACT QUANTIFICATION OF TYPE III COLLAGEN DEPOSITION DENSITY FROM PHOTOMICROGRAPH OF VAGINAL CONNECTIVE TISSUE By Muhammad Arfan NIM: 23218110 (Master’s Program in Electrical Engineering) The medical world has always relied on visualization in clinical trials to conclude a diagnosis. Unfortunately, this is often done manually with a relatively large sample of observations. For example, quantitative research involves hundreds of samples with complex research objects requiring precise accuracy that will take much time. As more samples need careful observation, the process becomes increasingly difficult. A study on the density of type III collagen was performed manually on 105 samples using the ImageJ program carried out on samples in the form of photomicrographs by changing the color of the type III collagen deposition in a binary image which resulted in an increase in deposition density in the form of a percentage of the stained area. This research will automatically quantify for a long process, and quantification does not require expertise in researching large samples and complex research objects. This study uses illumination correction to even out lighting and reduces noise on photomicrographs to facilitate cell identification. In addition, to increase the pixel intensity and the structure of the collagen fibers with line and tubeness feature enhancement. The CellProfiler module offers automatic quantification. The time it takes to process 105 images with CellProfiler takes 8 minutes 57 seconds or 5.1 seconds per image. While with ImageJ, it takes 1 minute 14 seconds per image or 129.5 minutes for 105 images (varies with researchers' level of accuracy). Image processing is 14.5 times faster by using CellProfiler. A comparison of the results of automatic calculations using CellProfiler and manual calculation of results using ImageJ was performed with Linear Regression which resulted in the value of R2 = 0.7786 and the function y = 0.9548x + 1.2197. The results of this comparison were positive with a strong correlation and indicated that automatic quantification with CellProfiler was successful in calculating type III collagen deposition and could be performed without special care from pathophysiologists. Keywords: prolaps, collagen, MSCs, image processing, automatic quantification.