Abstract | ||
---|---|---|
We present a solution for the goal of extracting a video from a single motion blurred image to sequentially reconstruct the clear views of a scene as beheld by the camera during the time of exposure. We first learn motion representation from sharp videos in an unsupervised manner through training of a convolutional recurrent video autoencoder network that performs a surrogate task of video reconstruction. Once trained, it is employed for guided training of a motion encoder for blurred images. This network extracts embedded motion information from the blurred image to generate a sharp video in conjunction with the trained recurrent video decoder. As an intermediate step, we also design an efficient architecture that enables real-time single image deblurring and outperforms competing methods across all factors: accuracy, speed, and compactness. Experiments on real scenes and standard datasets demonstrate the superiority of our framework over the state-of-the-art and its ability to generate a plausible sequence of temporally consistent sharp frames. |
Year | DOI | Venue |
---|---|---|
2018 | 10.1109/CVPR.2019.00699 | 2019 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2019) |
Field | DocType | Volume |
Video reconstruction,Autoencoder,Pattern recognition,Computer science,Artificial intelligence,Encoder,Video decoder | Journal | abs/1804.02913 |
ISSN | Citations | PageRank |
1063-6919 | 1 | 0.35 |
References | Authors | |
24 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Kuldeep Purohit | 1 | 13 | 7.65 |
Anshul Shah | 2 | 1 | 0.35 |
A. N. Rajagopalan | 3 | 1106 | 92.02 |