In search of a counter you can count on: relative efficacy of human visual and automated colony counting

Abstract

Colony quantification is essential in clinical and research settings as well as pedagogy at the college level. Human visual (HV) counting, the most common method, is time consuming and fraught with errors. The time, accuracy and precision of HV counting were compared to a high-end professional automated counter, an inexpensive phone application and a free phone application. Low cost benefits of increased speed and accuracy with automated counting are maximized when counting single round colonies; but much reduced if colonies have irregular morphology or demonstrate haemolysis.

Scan4CFU: Low-cost, open-source bacterial colony tracking over large areas and extended incubation times

“…Counting bacterial colonies that form on nutrient agar plates (i.e., plate counting) is a standard practice in microbiology in order to quantify the number of individual bacteria within a culture, and is based on reporting on the number of colony forming units (CFU)/mL of culture. However, this method is not designed to measure growth kinetics of bacteria as they form visible colonies [1]. There are multiple studies, however, that would benefit from the ability to monitor growth kinetics of bacterial colonies over prolonged time scales.…”

Evaluation of an Automated System for the Counting of Microbial Colonies

“…Moreover, high colony numbers may lead to false results because often only parts of a given plate are being counted. Automated colony counters have been developed for handling high plate volumes ( 6 ). However, the accuracy of counting should remain the highest priority ( 6 , 7 ).…”

“…Automated colony counters have been developed for handling high plate volumes ( 6 ). However, the accuracy of counting should remain the highest priority ( 6 , 7 ).…”

Experimental setup and image processing method for automatic enumeration of bacterial colonies on agar plates

“…In principle automated colony counting, as various publications suggest [1,[6][7][8][9], increases laboratory productivity considerably. Ideally there will be fewer faults, improved reproducibility, and the results become independent of the training level of the person who carries out the assessment.…”

“…Initial tests with the public domain image processing software package ImageJ showed that recording a digital image and CFU enumeration in software, executing all steps manually, required less time compared to manual counting for CFU numbers as low as 50 per plate provided that predefined regions of interest (defining the area to be processed) could be usedexperienced persons count approximately 2-3 colonies per second by marking & taking notes (see "Results" section below) [8]. At this stage it was estimated that fully automated processing of plate images might lead to a reduction in processing time by one order of magnitude compared to manual counting, with the additional benefit of simultaneous digital documentation.…”