Probabilistic reasoning in intelligent systems: networks of plausible inference
Probabilistic reasoning in intelligent systems: networks of plausible inference
A nonlinear filter for film restoration and other problems in image processing
CVGIP: Graphical Models and Image Processing
Conditional Random Fields: Probabilistic Models for Segmenting and Labeling Sequence Data
ICML '01 Proceedings of the Eighteenth International Conference on Machine Learning
Deterioration detection for digital film restoration
CVPR '97 Proceedings of the 1997 Conference on Computer Vision and Pattern Recognition (CVPR '97)
SIGGRAPH '84 Proceedings of the 11th annual conference on Computer graphics and interactive techniques
Bilateral Filtering for Gray and Color Images
ICCV '98 Proceedings of the Sixth International Conference on Computer Vision
Efficient Belief Propagation for Early Vision
International Journal of Computer Vision
Fast high quality interpolation of missing data in image sequences using a controlled pasting scheme
ICASSP '99 Proceedings of the Acoustics, Speech, and Signal Processing, 1999. on 1999 IEEE International Conference - Volume 06
MCMC for joint noise reduction and missing data treatment indegraded video
IEEE Transactions on Signal Processing
On missing data treatment for degraded video and film archives: a survey and a new Bayesian approach
IEEE Transactions on Image Processing
Interpolation of missing data in image sequences
IEEE Transactions on Image Processing
| Hi-index | 0.01 |
In this paper, we describe a new method for restoring digitized vintage video with film wear artifacts. Such artifacts result in partially or completely missing information. To maximize use of observed data, we cast the problem as that of recovering mattes of artifacts. More specifically, we extract the distributions of artifact color and its fractional (alpha) contribution to the frame. To account for spatial color discontinuity and pixel occlusion or disocclusion, we introduce the alpha-modulated bilateral filter. The problem is solved as a 3-D spatio-temporal conditional random field (CRF) with artifact color and (discretized) alpha as states. Inference is done through belief propagation. Results verify the effectiveness of our method. Furthermore, we can produce a synthetically generated vintage footage using extracted artifact information from actual vintage video.