MFSR: Light Field Images Spatial Super Resolution Model Integrated with Multiple Features

Light Field (LF) cameras can capture angular and spatial information simultaneously, making them suitable for a wide range of applications such as refocusing, disparity estimation, virtual reality. However, the limited resolution LF images hinders their applicability. In order to address this issue, we propose an end-to-end learning-based light field super-resolution (LFSR) model called MFSR, which integrates multiple features, including spatial, angular, epipolar plane (EPI), global features. These features are extracted separately from image then fused together obtain comprehensive feature using Feature Extract Block (FE Block) iteratively. Gradient loss is added into function ensure that MFSR has good performance with rich texture. Experimental results on synthetic real-world datasets demonstrate proposed method outperforms other state-of-the-art methods, peak signal-to-noise ratio (PSNR) improvement 0.208 dB 0.274 average 2× 4× tasks, structural similarity (SSIM) both improvements 0.01 average.