Collaborative research project / 2025
Seeing Spores
A collaborative food-science and engineering project that tested whether a small camera could help home fermenters recognize when tempeh begins to sporulate.
The problem
David and I began with a practical barrier to home fermentation: different substrates ferment at different rates, and visual monitoring requires frequent attention.
This project investigates a potential solution to this problem by building a device to visually monitor the development of mold growth during tempeh fermentation that can be used to detect the onset of spore growth without requiring involved and constant monitoring.
The result
After building the device and training substrate-specific image classifiers, we assessed both classification accuracy and how quickly the models identified visible spores.
The results of this project show that the trained CNN on device images is highly accurate (> 90%) in reporting the emergence of spores, and averages detection within 1 hour of emergence.
A possible next step
We treated the project as an initial prototype rather than a finished product.
Based on this project, a future iteration of this device could be built that could be easily integrated with other technology, like a phone app, to make at-home tempeh fermentation safer, easier, and more optimal.
