首先PP-YOLOE-l 在COCO数据集上达到了51.4mAP。相比较PP-YOLOv2提升1.9AP和13.35%的速度,相比较YOLOX提升1.3AP和24.96%的速度。PP-YOLOE中主要的改进点是:anchor-free,powerful backbone and neck,TAL动态label assign,ET-head。 该检测器的设计机制包括:Anchor free无锚盒机制,可扩展的backbone和neck,由CSPRepR...
论文标题:PP-YOLOE: An evolved version of YOLO 论文链接:arxiv.org/abs/2203.1625 代码地址:github.com/PaddlePaddle 后续更新:github.com/PaddlePaddle 一句话总结:PP-YOLOE是一种基于无锚框机制、高效任务对齐头和先进标签分配策略的高性能目标检测器,具有出色的速度和精度表现。 图1: PP-YOLOE 与其他最先进...
2) a PAN (path aggregation network) neck with an SPP layer and DropBlock, and 3) a lightweight IoU aware head. Similar to YOLOv3, PP-YOLOv2 only assigns one anchor box for each ground truth object. This mechanism however requires a number of extra hyperparamete...
上个月,百度飞桨团队开源了其最新SOTA通用检测模型——PP-YOLOE+,COCO数据集精度达54.7mAP,其l版本相比YOLOv7精度提升1.9%,V100端到端(包含前后处理)推理速度达42.2FPS,文章回顾请戳:
@article{xu2022pp, title={PP-YOLOE: An evolved version of YOLO}, author={Xu, Shangliang and Wang, Xinxin and Lv, Wenyu and Chang, Qinyao and Cui, Cheng and Deng, Kaipeng and Wang, Guanzhong and Dang, Qingqing and Wei, Shengyu and Du, Yuning and others}, journal={arXiv preprint ...
[3] Xu S,Wang X,Lv W,et al.PP-YOLOE: An evolved version of YOLO[J].arXiv preprint arXiv:2203.16250,2022. [4] Li Z,Hou B,Wu Z,et al.Fcosr:A simple anchor-free rotated detector for aerial object detection[J].arXiv preprint arXiv:2111.10780, 2021. [5] Llerena J M,Zeni L...
在DOTA 1.0数据集上,PP-YOLOE-R-l 和 PP-YOLOE-R-x 在单尺度训练和测试的情况下分别达到了 78.14mAP 和 78.28 mAP,这在单尺度评估下超越了几乎所有的旋转框检测模型。 通过多尺度训练和测试,PP-YOLOE-R-l 和 PP-YOLOE-R-x 的检测精度进一步提升至 80.02mAP 和 80.73 mAP,超越了所有的 Anchor-...
文章标题:PP-YOLOE: An evolved version of YOLO 文章链接:arxiv.org/abs/2203.1625 代码链接:github.com/PaddlePaddle 摘要: 百度基于其之前的SOTA模型PP-YOLOEv2,提出了工业界常用的新的演进版本PP-YOLOE,其有更高的检测精度和部署友好。该方法基于anchor-free的架构,使用强大的backbone和neck,引入了CSPRepResSta...
PP-YOLOE: An evolved version of YOLO[J]. arXiv preprint arXiv:2203.16250, 2022. [4] Li Z, Hou B, Wu Z, et al. Fcosr: A simple anchor-free rotated detector for aerial object detection[J]. arXiv preprint arXiv:2111.10780, 2021. [5] Llerena J M, Zeni L F, Kristen L N, ...
PP-YOLOE-x(CSPRepResNet-x, 640x640, single-scale ) box mAP 52.2 # 70 Compare AP5069.9# 39 Compare AP7556.5# 35 Compare APS33.3# 31 Compare APM56.3# 21 Compare APL66.4# 20 Compare Multiple Object TrackingCroHDPP-TrackingMOTA72.6# 1 ...