Reducing Label Noise in Anchor-Free Object Detection

BMVC 2020  ·  Nermin Samet, Samet Hicsonmez, Emre Akbas ·

Current anchor-free object detectors label all the features that spatially fall inside a predefined central region of a ground-truth box as positive. This approach causes label noise during training, since some of these positively labeled features may be on the background or an occluder object, or they are simply not discriminative features. In this paper, we propose a new labeling strategy aimed to reduce the label noise in anchor-free detectors. We sum-pool predictions stemming from individual features into a single prediction. This allows the model to reduce the contributions of non-discriminatory features during training. We develop a new one-stage, anchor-free object detector, PPDet, to employ this labeling strategy during training and a similar prediction pooling method during inference. On the COCO dataset, PPDet achieves the best performance among anchor-free top-down detectors and performs on-par with the other state-of-the-art methods. It also outperforms all major one-stage and two-stage methods in small object detection (${AP}_{S}$ $31.4$). Code is available at https://github.com/nerminsamet/ppdet

PDF Abstract

Datasets


Results from the Paper


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Object Detection COCO minival PPDet (ResNet-101-FPN) box AP 40.5 # 163
AP50 59.5 # 80
AP75 44.2 # 68
APS 25.4 # 49
APM 44.7 # 52
APL 52.3 # 67
Object Detection COCO test-dev PPDet (ResNeXt-101-FPN, multiscale) box mAP 46.3 # 122
AP50 64.8 # 80
AP75 51.6 # 61
APS 31.4 # 39
APM 49.9 # 59
APL 56.4 # 87
Hardware Burden None # 1
Operations per network pass None # 1

Methods


No methods listed for this paper. Add relevant methods here