mIoU: 77.5%Frame (fps): 102.6(1080Ti)
Papers with Code · CC BY-SA 4.0
Real-Time Semantic Segmentation
Semantic Segmentation is a computer vision task that involves assigning a semantic label to each pixel in an image. In Real-Time Semantic Segmentation, the goal is to perform this labeling quickly and accurately in real-time, allowing for the segmentation results to be used for tasks such as object recognition, scene understanding, and autonomous navigation.
<span style="color:grey; opacity: 0.6">( Image credit: TorchSeg )</span>
논문
mIoU: 76.8%Frame (fps): 97.0(1080Ti)
mIoU: 80.6%Time (ms): 32.2Frame (fps): 31.1(3090)
mIoU: 75.3%Time (ms): 21.1Frame (fps): 47.3
mIoU: 74.7%Time (ms): 15.2Frame (fps): 65.5
mIoU: 77.4%Time (ms): 9.8Frame (fps): 101.6(2080Ti)
mIoU: 70.7%Time (ms): 16Frame (fps): 63
mIoU: 71.8%Time (ms): 469Frame (fps): 2.1
mIoU: 75.9%Time (ms): 18.8
(1080Ti)Frame (fps): 53
(1080Ti)
mIoU: 75.8%Time (ms): 27.1Frame (fps): 36.9
mIoU: 75.5%Frame (fps): 39.9
mIoU: 71.3%Frame (fps): 18.4 (1080Ti)
mIoU: 70.6%Time (ms): 14Frame (fps): 71
mIoU: 78.3%Time (ms): 33Frame (fps): 30
mIoU: 73.8%Frame (fps): 48.6 (1080Ti)
mIoU: 80.4%Time (ms): 39.2Frame (fps): 25.7(1080Ti)
mIoU: 74.8%Time (ms): 16.9Frame (fps): 59.2
mIoU: 76.41%
mIoU: 74.9%Time (ms): 21Frame (fps): 47.6 (Titan X)