2013/01/16 by Sebastian Hitziger, Maureen Clerc, Hitziger, Sebastian +9
Computer Science · Earth and Planetary Sciences · Engineering · Neuroscience · Physics and Astronomy · #62-07 #Atomic and Subatomic Physics Research #Blind Source Separation Techniques #EEG and Brain-Computer Interfaces #FOS: Computer and information sciences #Ferroelectric and Negative Capacitance Devices #Geophysical and Geoelectrical Methods #Machine Learning (stat.ML) #Sparse and Compressive Sensing Techniques
paper · pdf · doi:10.48550/arxiv.1301.3611
openalex publication_date 2013/01/16 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Dictionary Learning has proven to be a powerful tool for many image\nprocessing tasks, where atoms are typically defined on small image patches. As\na drawback, the dictionary only encodes basic structures. In addition, this\napproach treats patches of different locations in one single set, which means a\nloss of information when features are well-aligned across signals. This is the\ncase, for instance, in multi-trial magneto- or electroencephalography (M/EEG).\nLearning the dictionary on the entire signals could make use of the alignement\nand reveal higher-level features. In this case, however, small missalignements\nor phase variations of features would not be compensated for. In this paper, we\npropose an extension to the common dictionary learning framework to overcome\nthese limitations by allowing atoms to adapt their position across signals. The\nmethod is validated on simulated and real neuroelectric data.\n