From AutoPIA
Jump to: navigation, search
AutoPIA splash screen

AutoPIA (Automatic Phantom Image Analysis) is a software for automatic recognition and analysis of test images of objects (phantoms) used to evaluate image quality. AutoPIA is able to quickly provide reproducible and objective quality indices for routine quality control processes aimed at optimizing the performance of an X-ray system.

AutoPIA automatically analyses a given digital image and quickly calculates objective, reliable, and reproducible indices to evaluate the performance of a radiological equipment. Then it stores the analysis results and provides control charts with baseline and threshold values in order to keep track of quality trends and abnormal conditions.

Table of contents


Before the advent of digital era in diagnostic imaging, routine quality controls were carried out by human observers that evaluated image quality by counting visible details and measuring optical density. The process was inevitably slow, affected by subjective criteria and variable over time, resulting in unreliable assessments with high costs.

With digital images, the process of analysis and evaluation can be automated, carrying along certain advantages:

  • economy: the computer is much faster and costs less than any human, manual analysis for complex phantoms takes about 20 minutes, if multiplied by the number of tests to be carried out for routine controls and the number of devices to control, the amount of work inevitably grows;
  • efficiency: an analysis performed by the computer is more sensitive and accurate; AutoPIA is able to calculate quality indices much more information-rich rather than the simple count of visible details, and this allows to highlight in advance situations where the performance of the equipment is slightly degrading, without having to wait for one or two details to become invisible;
  • objectivity: the analysis is not influenced by human choices; AutoPIA is completely independent; once the image has been selected and the phantom used has been specified, the analysis is fully automatic and ends with the measurement of quality indices, and evidence of which have exceeded the defined thresholds; furthermore, the indices can be inserted into the control charts to identify trends;
  • reproducibility: the same image will be evaluated in the same way, regardless of the person performing the analysis and her/his competence or experience.

AutoPIA methods and algorithms are based on advanced knowledge of medical physics and computer science, and this provides:

  • incontestable results: the quality indices used by AutoPIA have a solid scientific basis and were validated using appropriate methods;
  • transparency: AutoPIA declares the algorithms used to calculate the quality indices and lets check for every image and every detail, what are the areas used by the algorithms; no magic number, everything is documented directly on the image used;
  • ease: AutoPIA is very easy to use because it is completely automatic and conforms to the most common interfaces used for computer programs; it hides the complexity of analysis algorithms including artificial intelligence algorithms and advanced techniques of optimization;
  • compatibility: AutoPIA calculates the indices used by the manual analysis of images, allowing easy integration with the processes of quality control currently implemented;
  • reporting: for each quality index, AutoPIA generates a control chart that highlights the trends, with the possibility of establishing benchmarks that generate alarms if exceeded. In the case of scientific or statistical work of greater complexity, AutoPIA export analysis results into a format compatible with popular spreadsheets (MS Excel, OpenOffice Calc), therefore allowing the use of analysis results in complete autonomy.