Dynamic R-CNN: Towards High Quality Object Detection via Dynamic Training

13 Apr 2020Hongkai ZhangHong ChangBingpeng MaNaiyan WangXilin Chen

Although two-stage object detectors have continuously advanced the state-of-the-art performance in recent years, the training process itself is far from crystal. In this work, we first point out the inconsistency problem between the fixed network settings and the dynamic training procedure, which greatly affects the performance... (read more)

PDF Abstract

Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT BENCHMARK
Object Detection COCO test-dev Dynamic R-CNN (ResNet-101-DCN, multi-scale) box AP 50.1 # 11
AP50 68.3 # 15
AP75 55.6 # 12
APS 32.8 # 13
APM 53.0 # 12
APL 61.2 # 15

Methods used in the Paper