This deep learning-based study proposes a new bacterial colony-forming unit (CFU) detection system that could save 12 hours compared to the Environmental Protection Agency (EPA) methods.

Bacterial infections are one of the biggest challenges that both developing and developed nations struggle with, The USA alone spends over $4 billion annually to combat such diseases, and hence it is a necessity to continuously look for new ways to fight such infections, one of the most effective ways of tackling such conditions are early detection as early detection helps in containing the spread of bacteria. With traditional methods such as culturing bacteria on an agar plate or a liquid medium taking more than 24 hours, we must use modern technology to improve detection time.

While other methods take less time to produce results, they risk inaccurate results. However, in recent years, the development of Thin Film Transistors (TFT), with their high scalability, low heat generation, and low cost, has been applied to the field of biosensing built by researchers from the University of California, Los Angeles. However, in the previous results, it has not been able to differentiate between dead and alive CFUs ( Colony Forming Units). However, with new advancements in deep learning and AI, it has been possible to use this technology in biosensing.

The TFT-based portable system has significant advantages of being cost-effective and having ultra-large FOV(Field of View) of TFT image sensors, which can be further scaled up, achieving even lower costs with much larger FOVs based on,

The TFT-based method, though doesn’t have an instantaneous colony classification time but a slightly delayed time. Field-portable CFU imager comprises an illumination module( a tri-colour LED) and a TFT-based image sensor. The light from a tri-colour LED directly illuminates the samples and forms in-line holograms on the TFT image sensor. The TFT module has a controlling printed circuit board (PCB) that sends signals for image capture and an image sensor. For the illumination module, a tri-colour LED is controlled by a microcontroller through DC(constant) current LED driver and sequentially provides the red (620 nm), green (520 nm), and blue (420 nm) illumination beams.

The sample was placed inside an incubator to record the growth every five minutes; for each 5 minutes time interval, three images were collected sequentially using the TFT image sensor under red (620 nm) and green (520 nm), blue (460 nm) light. The wavelengths allow us to label each bacterium using colours. The detection rate for this particular case was defined as the number of colonies confirmed by the system vs the no.of colonies detected manually through traditional methods by an expert, > 90% detection rate was reported for 8 hours for E. coli, 9 hours for Citrobacter, and 7 hours 40 minutes for K. pneumoniae. Furthermore, a 100% detection rate was obtained for less than 10 hours of incubation for E. coli, 11 hrs Citrobacter, and 9 hrs 20 minutes for K. pneumoniae. The TFT-based CFU detection system achieved more than 12 hours of time-saving while displaying the maximum accuracy possible.

Overall, the Performance of the TFT-based CFU detection method is similar to other similar systems like the CMOS-based time-lapse imaging method10 in terms of the colony detection speed. But, because of the large pixel size (375 μm) and limited spatial resolution, In terms of Performance, the presented CFU detection system using TFT image sensor arrays provides a high throughput which is a cost-effective and easy-to-use solution for early detection and classification of bacterial colonies, which open up unique opportunities for microbiology instrumentation in the laboratory and field settings.

Through intensive testing, it is found that the methods applied in this research paper are not only practical but also faster, more cost-friendly, more portable, and more accurate. The equipment and technology are relatively accessible and cheap, which implies it is applied in remote parts of the world where these bacterial infections are common due to hygiene problems or unclean/contaminated food and water sources. For those areas, this technology can improve the living conditions reasonably significantly. 

Free-2 Min AI Newsletter Join 500,000+ AI Folks EmailEnter your email address Subscribe

Marktechpost is a California based AI News Platform providing easy-to-consume, byte size updates in machine learning, deep learning, and data science research

© 2021 Marktechpost LLC. All Rights Reserved. Made with ❤️ in California