To help demonstrate the research, the team took a series of clips from The Slow Mo Guys, a popular slow-motion based science and technology entertainment YouTube series created by Gavin Free, starring himself and his friend Daniel Gruchy, and made their videos even slower. “Our multi-frame approach consistently outperforms state-of-the-art single frame methods.” “Our method can generate multiple intermediate frames that are spatially and temporally coherent,” the researchers said.
The result can make videos shot at a lower frame rate look more fluid and less blurry. The team used a separate dataset to validate the accuracy of their system. Once trained, the convolutional neural network predicted the extra frames. Using NVIDIA Tesla V100 GPUs and cuDNN -accelerated PyTorch deep learning framework the team trained their system on over 11,000 videos of everyday and sports activities shot at 240 frames-per-second. With this new research, users can slow down their recordings after taking them. “While it is possible to take 240-frame-per-second videos with a cell phone, recording everything at high frame rates is impractical, as it requires large memories and is power-intensive for mobile devices,” the team explained. “There are many memorable moments in your life that you might want to record with a camera in slow-motion because they are hard to see clearly with your eyes: the first time a baby walks, a difficult skateboard trick, a dog catching a ball,” the researchers wrote in the research paper. The researchers will present their work at the annual Computer Vision and Pattern Recognition (CVPR) conference in Salt Lake City, Utah this week. Video contributed by Dr.Researchers from NVIDIA developed a deep learning -based system that can produce high-quality slow-motion videos from a 30-frame-per-second video, outperforming various state-of-the-art methods that aim to do the same. Note that the actual time between the appearance of the stepped leader to the return stroke is about 1/50 of a second.įigure 3 is the same as Figure 2, except played back at an even slower rate to show the apparent movement of the return stroke as the two channels are discharged. The first to make a connection discharges the entire channel. In this case, the tips of the stepped leaders are racing toward the ground. This video is less exposed than the previous example (which makes everything appear dimmer). Unlike the previous example, this video shows just two main channels of a stepped leader approaching ground. Only one return stroke was recorded with this flash.įigure 2 shows another stepped leader approaching the ground. Note that the actual time between the appearance of the stepped leader in the cloud until the return stroke is about 1/133 of a second. When one branch of the stepped leader makes a connection, a very bright return stroke surges upward through the channel. Played back at 30 frames/second, 1 second of recorded video lasts 2 minutes and 13 seconds.Īlthough somewhat overexposed, this video shows a highly branched step leader approaching the ground.
This video was recorded at 4000 frames/second. Figure 1 is an example of video from a high speed camera. When played back in slow motion, stepped leaders, upward leaders, streamers, dart leaders, and return strokes can be seen. These cameras are capable of capturing many stages of the lightning flash that the human eye cannot discern. Recent advances in video equipment allow videographers to capture high speed images of lightning.
0 kommentar(er)
