The code, particle_count_gui_tk_20x_adaptive.py, is used to count the amount of microparticle in a frame accurately by the image analysis and human hand- picking. The picture by optical microscope is transformed to the grayscale image first. Then, it is converted to the binary image by an adaptive threshold method. A median filter is used to remove the pepper noise and the hole in image is filled. The boundary in image is detected and classified by the area of each object in the frame. The range x of pixel to define the “single”, “double”, and “multiple” groups are “200 <= x < 624”, “624 <= x < 936”, and “936 < x” that these criteria can be changed based on the image’s resolution. Finally, each detected object can be checked to be single particle or not by the human hand-picking.