Dual-input deep learning system for microbial identification from blood agar plates

Takao Naito, Satomi Takei, Shigeki Misawa, Miyuki Kuribara, Yoko Tabe · 2026

Read Dual-input deep learning system for microbial identification from blood agar plates by Takao Naito, Satomi Takei, Shigeki Misawa, Miyuki Kuribara, Yoko Tabe (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/plos-pone-0353761.txt.

Description

Naito and colleagues, in a 2026 study, test whether a computer system can name harmful microbes from photos of their colonies on blood agar plates. Naming microbes by eye takes training and years of practice, and few labs have automated this step. Their system reads each plate in two ways at once.

One model studies single colonies cut out of the photo, while a second studies small tiles cut from the whole plate image. Each model was trained on thousands of images from hundreds of strains across ten species that matter in clinics. Used alone, each model scored well. Joined as a pair, they did better still on fresh strains they had not trained on, catching nearly all cases and rarely raising false alarms. The authors see the paired design as a practical help for busy microbiology labs.

Catalog record

CREATORTakao Naito, Satomi Takei, Shigeki Misawa, Miyuki Kuribara, Yoko Tabe
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE34251
LICENSECC_BY
RIGHTSCC BY
SOURCEPLOS

Source: https://doi.org/10.1371/journal.pone.0353761

Browse the full library · About the Tower of Babel