Tensor Completion Algorithms for Estimating Missing Values in Audio Domain
In this project, various tensor completion methodologies will be applied to a specific kind of data, i.e. sound spectro- gram. As it will be depicted in section 2, data completion has been a widely studied subject showing impressive per- formances to recover and fill missing data. However, most applications were made on images or movies. The structure of images is rather specific and the question of whether the methodologies could also be applied to drastically different data structures will be tested throughout this project.