PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
November 28, 2019IEEE Transactions on Pattern Analysis and Machine Intelligence1,889 citations

Cascade R-CNN: High Quality Object Detection and Instance Segmentation

View Full Paper
ZCZhaowei CaiNVNuno Vasconcelos

Key Points

Key points are not available for this paper at this time.

Abstract

In object detection, the intersection over union (IoU) threshold is frequently used to define positives/negatives. The threshold used to train a detector defines its quality. While the commonly used threshold of 0.5 leads to noisy (low-quality) detections, detection performance frequently degrades for larger thresholds. This paradox of high-quality detection has two causes: 1) overfitting, due to vanishing positive samples for large thresholds, and 2) inference-time quality mismatch between detector and test hypotheses. A multi-stage object detection architecture, the Cascade R-CNN, composed of a sequence of detectors trained with increasing IoU thresholds, is proposed to address these problems. The detectors are trained sequentially, using the output of a detector as training set for the next. This resampling progressively improves hypotheses quality, guaranteeing a positive training set of equivalent size for all detectors and minimizing overfitting. The same cascade is applied at inference, to eliminate quality mismatches between hypotheses and detectors. An implementation of the Cascade R-CNN without bells or whistles achieves state-of-the-art performance on the COCO dataset, and significantly improves high-quality detection on generic and specific object datasets, including VOC, KITTI, CityPerson, and WiderFace. Finally, the Cascade R-CNN is generalized to instance segmentation, with nontrivial improvements over the Mask R-CNN.

Ask AI
Helpful
Bookmark
Share
View Full Paper

Cite This Study

Cai et al. (2019) studied this question.

synapsesocial.com/papers/69d91acc7fca1f84ab684298https://doi.org/10.1109/tpami.2019.2956516
Ask AI
Helpful
Bookmark
Share
View Full Paper