Cascaded Partial Decoder for Fast and Accurate Salient Object Detection

CVPR 2019 Zhe WuLi SuQingming Huang

Existing state-of-the-art salient object detection networks rely on aggregating multi-level features of pre-trained convolutional neural networks (CNNs). Compared to high-level features, low-level features contribute less to performance but cost more computations because of their larger spatial resolutions... (read more)

PDF Abstract

Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT LEADERBOARD
Salient Object Detection DUT-OMRON CPD-R (ResNet50) MAE 0.056 # 2
F-measure 0.747 # 3
Salient Object Detection DUTS-test CPD-R (ResNet50) F-measure 80.5 # 1
MAE 0.043 # 1
Salient Object Detection ECSSD CPD-R (ResNet50) MAE 0.037 # 1
F-measure 0.917 # 3
Salient Object Detection HKU-IS CPD-R (ResNet50) MAE 0.034 # 3
F-measure 0.891 # 3
Salient Object Detection ISTD CPD Balanced Error Rate 6.76 # 1
Salient Object Detection PASCAL-S CPD-R (ResNet50) MAE 0.072 # 2
F-measure 0.824 # 3
Salient Object Detection SBU CPD Balanced Error Rate 4.19 # 1
Salient Object Detection UCF CPD Balanced Error Rate 7.21 # 1

Methods used in the Paper