sorry, we can't preview this file
LineDataNewNew128.mat (2.42 GB)
Augmented Training Data of H:Si(100)
dataset
posted on 2019-09-10, 09:43 authored by Oliver Gordon, Philip MoriartyThis dataset contains images of the H:Si(100) surface, taken over several years on a Scienta Omicron Variable Temperature STM based at the University of Nottingham, UK.
These images were converted from their proprietary z_mtrx format from scan sizes up to 512x512 resolution, and classified into six categories by hand. They were then augmented with randomly with crops, pans, zooms, rotations and random amounts of random noise, scaling down to 128x128.
