Comparison of detection results of different methods in FasterNext.

<p>(a) YOLOv5s. (b) YOLOv5s + FasterNext (Backbone). (c) YOLOv5s + FasterNext(Neck). (d) YOLOv5s + FasterNext(Backbone + Neck).</p>

محفوظ في:

التفاصيل البيبلوغرافية
المؤلف الرئيسي:	Junjie Lu (160350) (author)
مؤلفون آخرون:	Yuchen Zheng (4172806) (author), Liwei Guan (21514868) (author), Bing Lin (136655) (author), Wenzao Shi (22008619) (author), Junyan Zhang (532480) (author), Yunping Wu (9472035) (author)
منشور في:	2025
الموضوعات:	Space Science Biological Sciences not elsewhere classified Information Systems not elsewhere classified time fire detection small fire targets reduce resource consumption reduce computational cost limited computational resources improve detection performance hardware resource constraints enhance detection precision diou loss function yolo increases map paper proposes fcmi efficient deep learning div >< p deep learning significantly exacerbating rapid development practical value neck network lowers gflops lightweight design inference speed fasternext module experimental results edge scenarios edge devices based algorithm 40 %,
الوسوم:	إضافة وسم لا توجد وسوم, كن أول من يضع وسما على هذه التسجيلة!

مواد مشابهة