A research on anomaly detection models, Sep. 2020 – Feb. 2021

We were offered and assigned to a project in the aim of improving the pre-trained model, which infers defects and their pre-defined types possible within the given image.

While the project was subdivided into a multiple smaller sections each of which has its own objective, we were in charge of summarizing and coordinating state-of-the-arts techniques and methodologies so that the theories are in reach of the model developers within the application.

Some of the techniques have been taken into the model or at least have been attempted to do so, for which better precision arises eminent in the PoC; whereas some others turned out not to use in practice yet to give an insight for the instructive reference in another section of the project.