meituan / YOLOv6

YOLOv6: a single-stage object detection framework dedicated to industrial applications.
GNU General Public License v3.0
5.7k stars 1.03k forks source link

EIOU和SIOU #89

Closed sdreamforchen closed 2 years ago

sdreamforchen commented 2 years ago

您好,根据你们的测试试验,你们是不是测试过EIOU和SIOU呢? 这俩谁效果好些呢?

Chilicyy commented 2 years ago

Thanks for your attention! We did some experiments with different IOU loss like CIOU, EIOU and SIOU on YOLOv6s model, and the results showed that SIOU loss(43.1% mAP) > EIOU loss (42.9% mAP) > CIOU loss (42.8% mAP).

sdreamforchen commented 2 years ago

Thanks very much

---原始邮件--- 发件人: @.> 发送时间: 2022年6月27日(周一) 晚上7:15 收件人: @.>; 抄送: @.**@.>; 主题: Re: [meituan/YOLOv6] EIOU和SIOU (Issue #89)

Thanks for your attention! We did some experiments with different IOU loss like CIOU, EIOU and SIOU on YOLOv6s model, and the results showed that SIOU loss(43.1% mAP) > EIOU loss (42.9% mAP) > CIOU loss (42.8% mAP).

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>