Robust global and local color matching in stereoscopic omnidirectional content

Deep Color Mismatch Correction In Stereoscopic 3d Images

“…New studies attempt to combine global and local approaches [5,6]. They generally first focus on a global color correction to rectify the color values and bring them as close to the reference as possible.…”

“…The proposed method was compared with Dudek et al's local approach [5], and the global approaches of Grogan et al [9], Pitie et al [7], and Reinhard et al [12]. We did not compare against the most recent deep learning-based method [11] as neither code nor data are available.…”

“…The presence of color mismatch can reduce the quality of experience and cause problems when processing the S3D images [1]. For color mismatch correction, there are already several traditional methods [2][3][4][5][6][7][8][9], but not so many based on deep learning, which has already been successfully applied in other computer vision problems. Therefore, here we propose an effective solution based on a convolutional network.…”