Microbial Colony Image Enhancement Algorithm Based on Composite Adversarial D-DCGAN Neural Network
Wangxin Li, Mingfeng Ge, Jing Sun, Dongxin Shi, Chenyu Jiang, Costas C Papagiannitsis · 2026
Read Microbial Colony Image Enhancement Algorithm Based on Composite Adversarial D-DCGAN Neural Network by Wangxin Li, Mingfeng Ge, Jing Sun, Dongxin Shi, Chenyu Jiang, Costas C Papagiannitsis (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/PMC13609509.txt.
Description
Wangxin Li and five colleagues build a generative system to fake microbial colony photos for training detection software. Deep learning can count colonies on Petri dishes, but it needs many labeled images, and real colony photos take time to produce. Their Double Deep Convolutional GAN creates synthetic colonies in two moves, overlapping small single colonies into composites and generating complete plates. It then tunes how many synthetic images to add using several quality metrics at once.
Starting from 236 real samples, the optimized recipe added 1,320 synthetic images. The realism score improved from 38.9 to 37.1, lower being better, and downstream detectors validated the approach: colony detection accuracy reached 0.975. The authors present the composite adversarial design as a practical answer to data scarcity in microbiology imaging. Labs could expand small photo collections into training sets for automated colony analysis, cutting manual counting in medical and food-safety testing.
Catalog record
| CREATOR | Wangxin Li, Mingfeng Ge, Jing Sun, Dongxin Shi, Chenyu Jiang, Costas C Papagiannitsis |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 47 KB |
| LICENSE | https://creativecommons.org/licenses/by/4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |