Biomedical Image Processing & Cell Area Quantification

Key Engineering Contributions

1. 01
   Image Segmentation: Implemented Otsu's method to binarize grayscale TEM images, successfully isolating the viral particles from the background cellular environment.
2. 02
   Morphological Filtering: Applied custom morphological closing operations and strict geometric thresholds (area boundaries and eccentricity ≤ 0.75) to automatically filter out pixel noise, artifacts, and incomplete cell boundaries.
3. 03
   Quantitative Data Analysis: Calculated and converted the pixel-based area of the identified cells into physical dimensions, successfully extracting the centroid coordinates and determining a cumulative viral area of 41,151 nm2 across the analyzed sample.

Visual Documentation

Figure 1

01.png

Training images - pipeline of image processing the raw TEM images to masking the SARS-CoV-2 particles and determining their area.

Figure 2

02.png

Testing images - pipeline of image processing the raw TEM images to masking the SARS-CoV-2 particles and determining their area.