I. Introduction
Seismic data interpolation is indispensable for reconstructing and regularizing seismic data. We all know that incomplete seismic data directly affect the accuracy of structural imaging and reservoir prediction. During the past few decades, many methods for seismic data interpolation have been proposed to reconstruct the missing traces, such as rank-reduction-based methods [1]–[3], sparse mathematical transform-based methods [4], [5], frequency space (FX) prediction filter-based methods [6], wave-equation-based methods [7], and curvelet transform interpolation method [8], [9]. Support vector regression (SVR) [10] has also been explored for seismic data interpolation.