Dual Diversified Dynamical Gaussian Process Latent Variable Model for Video Repairing
Abstract of Dual Diversified Dynamical Gaussian Process Latent Variable
In this Dual Diversified Dynamical Gaussian Process Latent Variable Model for Video Repairing paper, we propose a dual diversified dynamical Gaussian process latent variable model (D3GPLVM) to tackle the video repairing issue. For preservation purposes, videos have to be conserved on media. However, storing on media, such as films and hard disks, can suffer from unexpected data loss, for instance, physical damage. So repairing of missing or damaged pixels is essential for better video maintenance. Most methods seek to fill in missing holes by synthesizing similar textures from local patches (the neighboring pixels), consecutive frames, or the whole video. However, these can introduce incorrect contexts, especially when the missing hole or number of damaged frames is large. Furthermore, simple texture synthesis can introduce artifacts in undamaged and recovered areas. To address aforementioned problems, we introduce two diversity encouraging priors to both of inducing points and latent variables for considering the variety in existing videos. In D3GPLVM, the inducing points constitute a smaller subset of observed data, while latent variables are a low-dimensional representation of observed data.
Conclusion
This Dual Diversified Dynamical Gaussian Process Latent Variable Model for Video Repairing paper presents D3 GPLVM for video repairing by simultaneously exploring dynamic and diversity properties under the GPLVM framework. Compared to DGPLVM, the diversity encouraging priors are essential in our model to extract more distinct features from the observed training frames. Meanwhile, our method has the inherent advantage of recovering incomplete frames with more complex sceneries since these types of video usually have more diverse characteristics that cannot be captured by traditional DGPLVM. umber of other methods on a movie dataset of 100 various video clips.
