NFFA online Scanning Electron Microscopy classifier automatically classify and tag your nano-images.
Some information coming from the instrument (Beamtime, Gun Vacuum, EHT, and so on) are automatically built into the images, but information about the target material is missing. Tagging the content of SEM images in a uniform way aims to produce Findable, Accessible, Interoperable and Retrievable data.
Different neural network architectures were trained and tested on a dataset of human-labelled SEM images. The best model achieved was used as the engine of the online analysis service we developed to automatically classify newly incoming images.
If you are not satisfied with the result, you can manually insert a new category.
Help us improving the performance of our neural network.
I am a development scientist at CNR/IOM with more than 15 years experience in scientific computing area and HPC computational e-infrastructures. I am currently coordinating a team that is maintaining cutting-edge HPC, GRID and CLOUD infrastructures and deliver high level computing services for CNR/IOM. From January 2014 I am coordinator of Master in High Performance Computing promoted by SISSA and ICTP (www.mhpc.it).