Auxiliary Detection Head for One-Stage Object Detection
Auxiliary Detection Head for One-Stage Object Detection
复制标题
用于一级物体检测的辅助检测头
DOI：
10.1109/access.2020.2992532
复制
发表时间：
2020
期刊：
影响因子：
3.9
通讯作者：
Fengzhong Qu
中科院分区：
文献类型：
--
作者：
Guozheng Jin;Rin-Ichiro Taniguchi;Fengzhong Qu
辅助分类器可以提高分类网络的性能。然而，在物体检测领域中尚未探索辅助检测头的效用。在本文中，我们提出了一个辅助检测头，以提高性能的一级对象检测器。与其他探测头类似，辅助探测头由分类子网和回归子网组成，本质上是两个卷积层。因此，辅助检测头在计算上是高效的。此外，辅助探测头实现了隐式两步级联回归。具体地，辅助检测头使用其输出盒作为进一步回归的锚。在辅助检测头内，对象定位的细化对应于将其输出框的位置调整到地面实况框，这有助于网络学习更鲁棒的特征。在推理时，可以在不影响主探测头性能的情况下去除辅助探测头，这得益于辅助探测头的独立性，并具有缩小模型规模和缩短推理时间的优点。在Pascal VOC和COCO数据集上对所提出的方法进行了评估。通过将辅助检测头合并到与主检测头并行的最先进的对象检测器中，我们在不同的基准上显示出其性能的一致改善，而在推理时没有引入额外的参数。
The auxiliary classifier can improve the performance of classification networks. However, the utility of the auxiliary detection head has not been explored in the object detection field. In this paper, we propose an auxiliary detection head to boost the performance of one-stage object detectors. Similar to other detection heads, the auxiliary detection head consists of a classification subnet and a regression subnet, which are essentially two convolution layers. Thus the auxiliary detection head is computationally efficient. Besides, the auxiliary detection head achieves implicit two-step cascaded regression. Specifically, the auxiliary detection head uses its output boxes as anchors for further regression. Within the auxiliary detection head, refinement of object localization corresponds to adjust the positions of its output boxes towards ground truth boxes, which helps the network learn more robust features. At inference, the auxiliary detection head can be removed without any adverse effect on the performance of the main detector head, which benefits from its independence and leads to two advantages: shrink the model size and shorten inference time. The proposed method is evaluated on Pascal VOC and COCO datasets. By incorporating the auxiliary detection head into a state-of-the-art object detector in parallel with the main detection head, we show consistent improvement over its performance on different benchmarks, whereas no extra parameters are introduced at inference time.
登录
查看更多内容
DOI：
10.1109/tpami.2018.2858826
发表时间：
2020-02-01
影响因子：
23.6
作者：
Lin, Tsung-Yi;Goyal, Priya;Dollar, Piotr
通讯作者：
Dollar, Piotr
影响因子：
19.5
作者：
Russakovsky, Olga;Deng, Jia;Fei-Fei, Li
通讯作者：
Fei-Fei, Li
DOI：
10.1007/978-3-030-01252-6_49
发表时间：
2017-12
期刊：
--
影响因子：
--
作者：
Hongyu Xu;Xutao Lv;Xiaoyu Wang;Zhou Ren;R. Chellappa
通讯作者：
Hongyu Xu;Xutao Lv;Xiaoyu Wang;Zhou Ren;R. Chellappa
DOI：
10.1109/tpami.2017.2745563
发表时间：
2016-10
影响因子：
23.6
作者：
Xingyu Zeng;Wanli Ouyang;Junjie Yan;Hongsheng Li;Tong Xiao;Kun Wang;Yu Liu;Yucong Zhou;Binh Yang;Zhe Wang;Hui Zhou;Xiaogang Wang
通讯作者：
Xingyu Zeng;Wanli Ouyang;Junjie Yan;Hongsheng Li;Tong Xiao;Kun Wang;Yu Liu;Yucong Zhou;Binh Yang;Zhe Wang;Hui Zhou;Xiaogang Wang
DOI：
10.1109/wacv45572.2020.9093364
发表时间：
2019-07
期刊：
2020 IEEE Winter Conference on Applications of Computer Vision (WACV)
影响因子：
--
作者：
Hoyong Jang;Sanghyun Woo;Philipp Benz;Jinsun Park;I. Kweon
通讯作者：
Hoyong Jang;Sanghyun Woo;Philipp Benz;Jinsun Park;I. Kweon