There is a lot of innovative work being conducted in industry, or in universities funded by industry. However, the aim of industrial innovation is not to propose novel scientific methods, but solutions to breakthrough two main dimensions: the performance (highest success with lowest critical error, resistance on specific constraints) and the user scope (application to a new domain, ability to adapt the application to another domain with a null or minimized effort). Hence, the nuance is that scientists target the methodology (how) when industrial fellows target the application (what).
Most applied research reuses, combines, tunes and trains existing methods with the advantage of accessing large corpuses of data. Industrial researchers either adapt or create methods with scientific innovations, but also with heuristics, workarounds, shortcuts, and whatever can pragmatically reach the target for an acceptable scope. Furthermore, we recognize that when industrial researchers create a real novel approach, the business strategy in the industrial competition prevents disclosure of the information. Hence, that applied research might not be able to be submitted at ICDAR.
However, all these works are indeed of interest to attendees of ICDAR, because experiments with large real-life datasets provide pragmatic feedback on technologies. They really demonstrate the state-of-the-art limitations, it infers a ranking of technology performances and, both workarounds or optimizations may point out some research perspectives. In addition, it highlights end-user problems, it shows industry orientations, i.e. it may infer academic perspectives. We do believe that, even if an application is not scientifically original or not clearly described, the result status and discussions around this method have a strong value for ICDAR attendees.
This workshop is an effort to make information about innovative applications available to the general ICDAR attendee, and to allow those researchers and engineers the opportunity to attend ICDAR. We hope it will also help build more interaction between industry and academia.