Computer vision (CV) is used nowadays in a wide range of real-world applications, from optical character recognition, autonomous driving and safety relevant vehicle functions, industrial inspection (rapid parts inspection for quality assurance), to medical imaging, fingerprint recognition and biometrics. Although a vast variety of literature covering evaluation techniques in subfields of the whole topic is available, still no study reports on testing a complete vision system, i.e., comprising hardware, software, data communication and control.

Obviously the high quality of CV applications has a great impact on their usability in real world scenarios. Hence beside traditional CV evaluation techniques – such as using test data sets as input and comparing the algorithm’s output against a manually established ground truth – we have to control the quality of the involved applications by means of applying a more generic evaluation strategy. In this context, QA activities like peer reviews, coding guidelines, usage of software quality tools (static and dynamic analysers) offer many benefits, from being able to track the CV project’s progress and estimate its relative complexity to helping us realize when we have achieved the desired state of quality.

The workshop welcomes contributions concerning QA activities applied in the CV domain, from CV specific evaluation techniques to more generic methods used both at software and hardware level in product development. We are particularly interested in original tools, languages and methods used for test data generation and test data validation in CV area, but also in testing tool evaluations based on vision application use cases.

Our aim is to bring together testing experts, computer vision developers, and industry practitioners in order to fill the gap between a typical quality assurance process and the current state of the evaluation methodology in different areas of computer vision discipline.