Cascade RetinaNet: Maintaining Consistency for Single-Stage Object Detection
arXiv:1907.06881
Abstract
Recent researches attempt to improve the detection performance by adopting the idea of cascade for single-stage detectors. In this paper, we analyze and discover that inconsistency is the major factor limiting the performance. The refined anchors are associated with the feature extracted from the previous location and the classifier is confused by misaligned classification and localization. Further, we point out two main designing rules for the cascade manner: improving consistency between classification confidence and localization performance, and maintaining feature consistency between different stages. A multistage object detector named Cas-RetinaNet, is then proposed for reducing the misalignments. It consists of sequential stages trained with increasing IoU thresholds for improving the correlation, and a novel Feature Consistency Module for mitigating the feature inconsistency. Experiments show that our proposed Cas-RetinaNet achieves stable performance gains across different models and input scales. Specifically, our method improves RetinaNet from 39.1 AP to 41.1 AP on the challenging MS COCO dataset without any bells or whistles.
BMVC 2019
References in corpus (8)
- Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
- YOLOv3: An Incremental Improvement
- DSSD : Deconvolutional Single Shot Detector
- Beyond Skip Connections: Top-Down Modulation for Object Detection
- Scale-Aware Trident Networks for Object Detection
- WIDER Face and Pedestrian Challenge 2018: Methods and Results
- Consistent Optimization for Single-Shot Object Detection
- Cascade Region Proposal and Global Context for Deep Object Detection
Cited by in corpus (5)
- R3Det: Refined Single-Stage Detector with Feature Refinement for Rotating Object
- Dynamic R-CNN: Towards High Quality Object Detection via Dynamic Training
- Objects in Semantic Topology
- Learning from Noisy Anchors for One-stage Object Detection
- CPM R-CNN: Calibrating Point-guided Misalignment in Object Detection