Image-Based Culture Plate Readers and Microbiological Testing of Mixtures - is It Necessary?

As the use of imaging devices and artificial intelligence (AI) gain traction in culture plate reading, the question of performance on detection of mixtures of organisms often arises. This is typically borne through traditional thoughts of validating alternative quantitative methods for micro-organisms to ensure methods do not demonstrate any interference or bias towards organism recovery or detection. This approach, however, may not be applicable to newer technologies. In the case of the APAS Independence, an AI-based culture plate reading system, testing of bacterial mixtures is not necessary. The system is designed to detect both bacterial and mold colonies, with each colony being classified as a result of pixel level detection in an independent and uninfluenced manner. The pixel result is entirely based on the way the algorithm is developed, which remains agnostic to any species identification within bacterial or mold groups themselves and only requires differentiation between these groups at a high level. This poster will provide insights into imaging and AI-analysis tools to demonstrate how mixtures of organisms are detected, highlighting the independent nature of colony assignment. If the segregation of mold and bacterial colonies is important, additional specificity testing may be considered.