Antibiogramj: A tool for analysing images from disk diffusion tests

  1. Alonso, C.A. 1
  2. Domínguez, C. 1
  3. Heras, J. 1
  4. Mata, E. 1
  5. Pascual, V. 1
  6. Torres, C. 1
  7. Zarazaga, M. 1
  1. 1 Universidad de La Rioja
    info

    Universidad de La Rioja

    Logroño, España

    ROR https://ror.org/0553yr311

Revista:
Computer Methods and Programs in Biomedicine

ISSN: 0169-2607

Año de publicación: 2017

Volumen: 143

Páginas: 159-169

Tipo: Artículo

DOI: 10.1016/J.CMPB.2017.03.010 SCOPUS: 2-s2.0-85015400447 WoS: WOS:000400531900017 GOOGLE SCHOLAR

Otras publicaciones en: Computer Methods and Programs in Biomedicine

Resumen

Background and objectives Disk diffusion testing, known as antibiogram, is widely applied in microbiology to determine the antimicrobial susceptibility of microorganisms. The measurement of the diameter of the zone of growth inhibition of microorganisms around the antimicrobial disks in the antibiogram is frequently performed manually by specialists using a ruler. This is a time-consuming and error-prone task that might be simplified using automated or semi-automated inhibition zone readers. However, most readers are usually expensive instruments with embedded software that require significant changes in laboratory design and workflow. Methods Based on the workflow employed by specialists to determine the antimicrobial susceptibility of microorganisms, we have designed a software tool that, from images of disk diffusion tests, semi-automatises the process. Standard computer vision techniques are employed to achieve such an automatisation. Results We present AntibiogramJ, a user-friendly and open-source software tool to semi-automatically determine, measure and categorise inhibition zones of images from disk diffusion tests. AntibiogramJ is implemented in Java and deals with images captured with any device that incorporates a camera, including digital cameras and mobile phones. The fully automatic procedure of AntibiogramJ for measuring inhibition zones achieves an overall agreement of 87% with an expert microbiologist; moreover, AntibiogramJ includes features to easily detect when the automatic reading is not correct and fix it manually to obtain the correct result. Conclusions AntibiogramJ is a user-friendly, platform-independent, open-source, and free tool that, up to the best of our knowledge, is the most complete software tool for antibiogram analysis without requiring any investment in new equipment or changes in the laboratory. © 2017 Elsevier B.V.