RVOS: End-to-End Recurrent Network for Video Object Segmentation

CVPR 2019 Carles VenturaMiriam BellverAndreu GirbauAmaia SalvadorFerran MarquesXavier Giro-i-Nieto

Multiple object video object segmentation is a challenging task, specially for the zero-shot case, when no object mask is given at the initial frame and the model has to find the objects to be segmented along the sequence. In our work, we propose a Recurrent network for multiple object Video Object Segmentation (RVOS) that is fully end-to-end trainable... (read more)

PDF Abstract CVPR 2019 PDF CVPR 2019 Abstract

Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT BENCHMARK
Semi-Supervised Video Object Segmentation DAVIS 2017 (test-dev) RVOS J&F 50.3 # 16
Jaccard (Mean) 47.9 # 15
Jaccard (Recall) 54.4 # 11
Jaccard (Decay) 35.7 # 16
F-measure (Mean) 52.6 # 17
F-measure (Recall) 61.7 # 13
F-measure (Decay) 36.7 # 16
Unsupervised Video Object Segmentation DAVIS 2017 (test-dev) RVOS J&F 22.5 # 5
Jaccard (Mean) 17.7 # 5
Jaccard (Recall) 16.2 # 5
Jaccard (Decay) 1.6 # 2
F-measure (Mean) 27.3 # 5
F-measure (Recall) 24.8 # 5
F-measure (Decay) 1.8 # 2
Semi-Supervised Video Object Segmentation DAVIS 2017 (val) RVOS Jaccard (Mean) 57.5 # 17
Jaccard (Recall) 65.2 # 15
Jaccard (Decay) 24.9 # 17
F-measure (Mean) 63.6 # 17
F-measure (Recall) 73.2 # 15
F-measure (Decay) 28.2 # 19
J&F 60.55 # 18
Unsupervised Video Object Segmentation DAVIS 2017 (val) RVOS J&F 41.2 # 7
Jaccard (Mean) 36.8 # 7
Jaccard (Recall) 40.2 # 7
Jaccard (Decay) 0.5 # 4
F-measure (Mean) 45.7 # 7
F-measure (Recall) 46.4 # 7
F-measure (Decay) 1.7 # 3
Youtube-VOS YouTube-VOS RVOS-Mask-ST+ F-Measure (Seen) 67.2 # 1
F-Measure (Unseen) 51 # 2
Jaccard (Seen) 63.6 # 1
Jaccard (Unseen) 45.5 # 2

Methods used in the Paper


METHOD TYPE
🤖 No Methods Found Help the community by adding them if they're not listed; e.g. Deep Residual Learning for Image Recognition uses ResNet