Deformable Convolutional Networks

Convolutional neural networks (CNNs) are inherently limited to model geometric transformations due to the fixed geometric structures in its building modules. In this work, we introduce two new modules to enhance the transformation modeling capacity of CNNs, namely, deformable convolution and deformable RoI pooling... (read more)

PDF Abstract ICCV 2017 PDF ICCV 2017 Abstract

Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT BENCHMARK
Object Detection COCO test-dev DeformConv-R-FCN (Aligned-Inception-ResNet) box AP 37.5 # 76
AP50 58.0 # 71
APS 19.4 # 75
APM 40.1 # 71
APL 52.5 # 57

Methods used in the Paper