Abstract | ||
---|---|---|
We propose an inertial-aided deblurring method that incorporates gyroscope measurements into a convolutional neural network (CNN). With the help of inertial measurements, it can handle extremely strong and spatially-variant motion blur. At the same time, the image data is used to overcome the limitations of gyro-based blur estimation. To train our network, we also introduce a novel way of generating realistic training data using the gyroscope. The evaluation shows a clear improvement in visual quality over the state-of-the-art while achieving real-time performance. Furthermore, the method is shown to improve the performance of existing feature detectors and descriptors against the motion blur. |
Year | Venue | Field |
---|---|---|
2018 | arXiv: Computer Vision and Pattern Recognition | Inertial frame of reference,Training set,Computer vision,Gyroscope,Deblurring,Feature detection,Convolutional neural network,Computer science,Motion blur,Artificial intelligence |
DocType | Volume | Citations |
Journal | abs/1810.00986 | 0 |
PageRank | References | Authors |
0.34 | 11 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Janne Mustaniemi | 1 | 2 | 2.41 |
Juho Kannala | 2 | 867 | 60.91 |
Simo Särkkä | 3 | 623 | 66.52 |
Jiri Matas | 4 | 335 | 35.85 |
Janne Heikkilä | 5 | 2163 | 160.55 |