不同检测模型预处理中的归一化操作对比 | YOLO 系列 | 【归一化操作归纳整理】
【摘要】
文章目录
YOLO 系列论文不同【检测模型】预处理中的归一化操作对比YOLO系列网络结构YOLOYolov2Yolov3Yolov4Yolov5
YOLO 系列论文
...
YOLO 系列论文
论文地址
论文翻译
论文解析
代码
- https://github.com/pjreddie/darknet 主页 – YOLOv4 可达
- https://github.com/ultralytics/yolov3
- https://github.com/ultralytics/yolov5
总结
- YOLO YOLOv2 YOLOv3 YOLOv4 都首先 基于 darknet 实现
- 然而,当前热度较高的版本通常还是 pytorch 实现,例如 yolov5
不同【检测模型】预处理中的归一化操作对比
# 如下归一化的模型有:
yolov2
const float mean_vals[3] = {1.0f, 1.0f, 1.0f};
const float norm_vals[3] = {0.007843f, 0.007843f, 0.007843f};
yolov3
const float mean_vals[3] = {127.5f, 127.5f, 127.5f};
const float norm_vals[3] = {0.007843f, 0.007843f, 0.007843f};
yolov4
const float mean_vals[3] = {0, 0, 0};
const float norm_vals[3] = {1 / 255.f, 1 / 255.f, 1 / 255.f};
in.substract_mean_normalize(mean_vals, norm_vals);
yolov5
const float norm_vals[3] = {1 / 255.f, 1 / 255.f, 1 / 255.f};
in_pad.substract_mean_normalize(0, norm_vals);
squeezenetssd squeezenet squeezenet_c_api
const float mean_vals[3] = {104.f, 117.f, 123.f};
in.substract_mean_normalize(mean_vals, 0);
simplepose
const float mean_vals[3] = {0.485f * 255.f, 0.456f * 255.f, 0.406f * 255.f};
const float norm_vals[3] = {1 / 0.229f / 255.f, 1 / 0.224f / 255.f, 1 / 0.225f / 255.f};
shufflenetv2
const float norm_vals[3] = {1 / 255.f, 1 / 255.f, 1 / 255.f};
in.substract_mean_normalize(0, norm_vals);
scrfd
const float mean_vals[3] = {127.5f, 127.5f, 127.5f};
const float norm_vals[3] = {1 / 128.f, 1 / 128.f, 1 / 128.f};
rfcn fasterrcnn
const float mean_vals[3] = {102.9801f, 115.9465f, 122.7717f};
in.substract_mean_normalize(mean_vals, 0);
nanodet
const float mean_vals[3] = {103.53f, 116.28f, 123.675f};
const float norm_vals[3] = {0.017429f, 0.017507f, 0.017125f};
mobilenetv3ssdlite
const float mean_vals[3] = {123.675f, 116.28f, 103.53f};
const float norm_vals[3] = {1.0f, 1.0f, 1.0f};
mobilenetv2ssdlite
const float mean_vals[3] = {127.5f, 127.5f, 127.5f};
const float norm_vals[3] = {1.0 / 127.5, 1.0 / 127.5, 1.0 / 127.5};
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
- 13
- 14
- 15
- 16
- 17
- 18
- 19
- 20
- 21
- 22
- 23
- 24
- 25
- 26
- 27
- 28
- 29
- 30
- 31
- 32
- 33
- 34
- 35
- 36
- 37
- 38
- 39
- 40
- 41
- 42
- 43
- 44
- 45
- 46
- 47
- 48
- 49
- 50
- 51
- 52
- 53
- 54
- 55
- 56
- 57
- 58
- 59
- 60
- 61
- 62
- 63
- 64
- 65
- 66
- 67
YOLO系列网络结构
YOLO
Yolov2
Yolov3
Yolov4
Yolov5
文章来源: positive.blog.csdn.net,作者:墨理学AI,版权归原作者所有,如需转载,请联系作者。
原文链接:positive.blog.csdn.net/article/details/119451088
【版权声明】本文为华为云社区用户转载文章,如果您发现本社区中有涉嫌抄袭的内容,欢迎发送邮件进行举报,并提供相关证据,一经查实,本社区将立刻删除涉嫌侵权内容,举报邮箱:
cloudbbs@huaweicloud.com
- 点赞
- 收藏
- 关注作者
评论(0)