Enhancing 360-Degree Video Streaming via Selective Inpainting for Bandwidth Optimization

360-degree video streaming is an emerging technology that immerses viewers in the dynamic environment of a video. However, streaming such videos over wireless networks faces challenges due to their substantial data demands. This research introduces an innovative approach aimed at mitigating the data requirements of 360-degree video streaming. Our proposed method leverages video inpainting to reconstruct select segments of the 360-degree video. To ensure a seamless integration of video inpainting without compromising the user's quality of experience, we devise a mechanism to identify tiles suitable for reconstruction with minimal impact on video quality. Empirical findings demonstrate that the proposed solution markedly reduces the data demand of 360-degree video streaming while maintaining satisfactory video quality.