DSBEC

The data set consists of 6257 labeled images of Bose-Einstein condensates (BECs) with and without solitonic excitations, including kink solitons and solitonic vortices. Each element of the data set contains a masked image (132x164 pixels) of 2D atomic density used to train the machine learning model used in the paper "Machine-learning enhanced dark soliton detection in Bose-Einstein condensates," (https://arxiv.org/abs/2101.05404), and a label indicating the class a given image belongs to (0 indicates no solitons, 1 indicates a single soliton, and 2 indicates other excitations). The data structure file and project description are included with the data. This data set was used to train a deep convolutional neural network to automatically recognize whether or not a lone dark soliton has been created in BECs that was then implemented within an automated soliton detection and positioning system (see https://arxiv.org/abs/2101.05404 for details).