Ergonomic image and video annotator
Scalable datastoring and processing capacities practically make machine learning based object and event detection on images and videos available for any data scientists.The importance of a good quality labeled image database is often forgotten. This requires the organisation of data collection and the ergonomic use of so-called object annotation (selecting, tagging) software.In case of the latter one efficiency, scalability and easy operability are the most important factors. For the time being, there is no "final solution", which would entirely meet all the requirements.
The aim of our solution is to set up a web-based service-oriented image object annotator environment, which makes it easier to label objects on images using smart structuring of the image sets and does not require professional knowledge in development. Our solution can be utilized in the field of ADAS (automated driving assistance), UAV-RS (unmanned aerial vehicle – remote sensing), product quality monitoring, or data collection via social media. Considering other solutions which are presently available the annotation process is way faster and more consistent results can be achieved. This tool reduces the development cycle of the industrial solutions created in the field of computer vision and will affect the productivity of several other industries (automotive industry, traffic, health care, energetics, etc.).