Time-Mapping Using Space-Time Saliency

CVPR 2014  ·  Feng Zhou, Sing Bing Kang, Michael F. Cohen ·

We describe a new approach for generating regular-speed, low-frame-rate (LFR) video from a high-frame-rate (HFR) input while preserving the important moments in the original. We call this time-mapping, a time-based analogy to high dynamic range to low dynamic range spatial tone-mapping. Our approach makes these contributions: (1) a robust space-time saliency method for evaluating visual importance, (2) a re-timing technique to temporally resample based on frame importance, and (3) temporal filters to enhance the rendering of salient motion. Results of our space-time saliency method on a benchmark dataset show it is state-of-the-art. In addition, the benefits of our approach to HFR-to-LFR time-mapping over more direct methods are demonstrated in a user study.

PDF Abstract

Results from the Paper


Task Dataset Model Metric Name Metric Value Global Rank Uses Extra
Training Data
Benchmark
Video Salient Object Detection DAVIS-2016 TIMP S-Measure 0.574 # 9
MAX E-MEASURE 0.680 # 8
AVERAGE MAE 0.185 # 1
Video Salient Object Detection DAVSOD-Difficult20 TIMP S-Measure 0.530 # 6
max E-measure 0.665 # 6
Average MAE 0.190 # 6
Video Salient Object Detection DAVSOD-easy35 TIMP S-Measure 0.534 # 7
max E-Measure 0.582 # 8
Average MAE 0.206 # 6
Video Salient Object Detection DAVSOD-Normal25 TIMP S-Measure 0.503 # 7
max E-measure 0.616 # 5
Average MAE 0.245 # 6
Video Salient Object Detection FBMS-59 TIMP S-Measure 0.576 # 15
AVERAGE MAE 0.192 # 14
MAX F-MEASURE 0.465 # 15
Video Salient Object Detection UVSD TIMP S-Measure 0.541 # 8
max E-measure 0.662 # 8
Average MAE 0.171 # 8
Video Salient Object Detection ViSal TIMP S-Measure 0.612 # 9
max E-measure 0.743 # 9
Average MAE 0.170 # 9
Video Salient Object Detection VOS-T TIMP S-Measure 0.546 # 9
max E-measure 0.640 # 9
Average MAE 0.192 # 9

Methods


No methods listed for this paper. Add relevant methods here