Algorithm 1. Pseudocode of the RBC single-cell extraction and resizing technique

RBC_Extraction(image_path, output_dir)   load image from image_path   apply mean-shift filtering to image  → shifted   convert shifted image to grayscale  → gray   apply Otsu thresholding to gray   → thresh   find contours from thresh      → cnts   for each contour c in cnts do     crop image and mask around contour → image_crop, mask_crop     if mask_crop is valid then       check if cell touches border:         if true:           save touching cell to “touching” folder         else:           extract RBC from mask           determine RBC size:           if size ≤ 16 px: overlay to 32 × 32 and save as “small”           else if size ≤ 32 px: overlay to 32 × 32 and save as “32 size”           else if size ≤ 128 px: overlay to 128 × 128 and save as “128 size”           else if size ≤ 256 px: overlay to 256 × 256 and save as “256 size”           else if size ≤ 512 px: overlay to 512 × 512 and save as “512 size”           else if size ≤ 1024 px: overlay to 1024 × 1024 and save as “1024 size”           else: save as “oversize”     else:   end for   save processing results (original, filtered, gray, mask, contours)   return RBC_dataset_summary