基于YOLOv5开发构建手部X光骨骼检测识别分析系统

最新推荐文章于 2025-02-10 00:15:00 发布

Together_CZ

最新推荐文章于 2025-02-10 00:15:00 发布

阅读量805

点赞数

文章标签： YOLO

本文链接：https://blog.csdn.net/Together_CZ/article/details/129627867

版权

本文介绍了在一个X光场景中进行目标检测的项目，使用了YOLO和VOC两种格式的标注文件。选择了轻量级的YOLO模型进行训练，并展示了训练日志、混淆矩阵、F1值曲线和PR曲线等评估指标，以及可视化推理的示例。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

X光场景的目标检测之前的项目中做的比较少，今天工作结束后正好有点时间就想着把之前遗留的数据集应用起来，开发一个实际的项目，首先看下效果图：

简单看下数据集如下：

YOLO格式标注文件如下：

实例标注内容如下所示：

3 0.746044 0.293727 0.109177 0.085645
3 0.609968 0.217732 0.109177 0.083233
3 0.480222 0.212907 0.093354 0.08082
3 0.336234 0.256333 0.102848 0.083233
0 0.392405 0.183353 0.098101 0.077201
0 0.560918 0.116405 0.112342 0.095296
0 0.691456 0.115199 0.132911 0.109771
0 0.829114 0.227382 0.148734 0.095296
0 0.839399 0.486128 0.162975 0.127865
4 0.643196 0.369723 0.128165 0.095296
4 0.530854 0.331122 0.109177 0.095296
4 0.421677 0.326297 0.093354 0.092883
4 0.322785 0.351628 0.098101 0.083233
2 0.560918 0.588058 0.143987 0.109771
4 0.711234 0.534379 0.112342 0.101327
1 0.318829 0.442702 0.09019 0.139928
1 0.394778 0.426417 0.087025 0.119421
1 0.477848 0.415561 0.101266 0.104946
1 0.571994 0.442702 0.102848 0.098914
5 0.425633 0.711701 0.161392 0.190591
6 0.318829 0.716526 0.106013 0.15199

VOC格式标注文件如下所示：

实例标注内容如下所示：

<?xml version="1.0" ?>
<annotation>
<folder>JPEGImages</folder>
<filename>0bb6e995-3a11-42a5-990c-d041a4b0e1d6.png</filename>
<path>0bb6e995-3a11-42a5-990c-d041a4b0e1d6.png</path>
<source>
    <database>Unknown</database>
</source>
<size>
    <width>1514</width>
    <height>2044</height>
    <depth>3</depth>
</size>

<segmented>0</segmented>
    <object>
    <name>MiddlePhalanx</name>
    <pose>Unspecified</pose>
    <truncated>0</truncated>
    <difficult>0</difficult>
    <bndbox>
        <xmin>832</xmin>
        <ymin>589</ymin>
        <xmax>932</xmax>
        <ymax>693</ymax>
    </bndbox>
</object>
    <object>
    <name>MiddlePhalanx</name>
    <pose>Unspecified</pose>
    <truncated>0</truncated>
    <difficult>0</difficult>
    <bndbox>
        <xmin>728</xmin>
        <ymin>545</ymin>
        <xmax>825</xmax>
        <ymax>645</ymax>
    </bndbox>
</object>
    <object>
    <name>MiddlePhalanx</name>
    <pose>Unspecified</pose>
    <truncated>0</truncated>
    <difficult>0</difficult>
    <bndbox>
        <xmin>625</xmin>
        <ymin>608</ymin>
        <xmax>710</xmax>
        <ymax>693</ymax>
    </bndbox>
</object>
    <object>
    <name>MiddlePhalanx</name>
    <pose>Unspecified</pose>
    <truncated>0</truncated>
    <difficult>0</difficult>
    <bndbox>
        <xmin>525</xmin>
        <ymin>700</ymin>
        <xmax>621</xmax>
        <ymax>786</ymax>
    </bndbox>
</object>
    <object>
    <name>DistalPhalanx</name>
    <pose>Unspecified</pose>
    <truncated>0</truncated>
    <difficult>0</difficult>
    <bndbox>
        <xmin>528</xmin>
        <ymin>530</ymin>
        <xmax>625</xmax>
        <ymax>667</ymax>
    </bndbox>
</object>
    <object>
    <name>DistalPhalanx</name>
    <pose>Unspecified</pose>
    <truncated>0</truncated>
    <difficult>0</difficult>
    <bndbox>
        <xmin>628</xmin>
        <ymin>386</ymin>
        <xmax>721</xmax>
        <ymax>545</ymax>
    </bndbox>
</object>
    <object>
    <name>DistalPhalanx</name>
    <pose>Unspecified</pose>
    <truncated>0</truncated>
    <difficult>0</difficult>
    <bndbox>
        <xmin>713</xmin>
        <ymin>334</ymin>
        <xmax>810</xmax>
        <ymax>482</ymax>
    </bndbox>
</object>
    <object>
    <name>DistalPhalanx</name>
    <pose>Unspecified</pose>
    <truncated>0</truncated>
    <difficult>0</difficult>
    <bndbox>
        <xmin>810</xmin>
        <ymin>426</ymin>
        <xmax>906</xmax>
        <ymax>552</ymax>
    </bndbox>
</object>
    <object>
    <name>DistalPhalanx</name>
    <pose>Unspecified</pose>
    <truncated>0</truncated>
    <difficult>0</difficult>
    <bndbox>
        <xmin>1021</xmin>
        <ymin>704</ymin>
        <xmax>1136</xmax>
        <ymax>893</ymax>
    </bndbox>
</object>
    <object>
    <name>ProximalPhalanx</name>
    <pose>Unspecified</pose>
    <truncated>0</truncated>
    <difficult>0</difficult>
    <bndbox>
        <xmin>547</xmin>
        <ymin>823</ymin>
        <xmax>643</xmax>
        <ymax>949</ymax>
    </bndbox>
</object>
    <object>
    <name>ProximalPhalanx</name>
    <pose>Unspecified</pose>
    <truncated>0</truncated>
    <difficult>0</difficult>
    <bndbox>
        <xmin>639</xmin>
        <ymin>763</ymin>
        <xmax>728</xmax>
        <ymax>904</ymax>
    </bndbox>
</object>
    <object>
    <name>ProximalPhalanx</name>
    <pose>Unspecified</pose>
    <truncated>0</truncated>
    <difficult>0</difficult>
    <bndbox>
        <xmin>736</xmin>
        <ymin>737</ymin>
        <xmax>836</xmax>
        <ymax>867</ymax>
    </bndbox>
</object>
    <object>
    <name>ProximalPhalanx</name>
    <pose>Unspecified</pose>
    <truncated>0</truncated>
    <difficult>0</difficult>
    <bndbox>
        <xmin>862</xmin>
        <ymin>741</ymin>
        <xmax>962</xmax>
        <ymax>863</ymax>
    </bndbox>
</object>
    <object>
    <name>ProximalPhalanx</name>
    <pose>Unspecified</pose>
    <truncated>0</truncated>
    <difficult>0</difficult>
    <bndbox>
        <xmin>961</xmin>
        <ymin>907</ymin>
        <xmax>1091</xmax>
        <ymax>1033</ymax>
    </bndbox>
</object>
    <object>
    <name>MCP</name>
    <pose>Unspecified</pose>
    <truncated>0</truncated>
    <difficult>0</difficult>
    <bndbox>
        <xmin>558</xmin>
        <ymin>934</ymin>
        <xmax>658</xmax>
        <ymax>1093</ymax>
    </bndbox>
</object>
    <object>
    <name>MCP</name>
    <pose>Unspecified</pose>
    <truncated>0</truncated>
    <difficult>0</difficult>
    <bndbox>
        <xmin>643</xmin>
        <ymin>893</ymin>
        <xmax>739</xmax>
        <ymax>1056</ymax>
    </bndbox>
</object>
    <object>
    <name>MCP</name>
    <pose>Unspecified</pose>
    <truncated>0</truncated>
    <difficult>0</difficult>
    <bndbox>
        <xmin>739</xmin>
        <ymin>849</ymin>
        <xmax>825</xmax>
        <ymax>1015</ymax>
    </bndbox>
</object>
    <object>
    <name>MCP</name>
    <pose>Unspecified</pose>
    <truncated>0</truncated>
    <difficult>0</difficult>
    <bndbox>
        <xmin>843</xmin>
        <ymin>852</ymin>
        <xmax>950</xmax>
        <ymax>1019</ymax>
    </bndbox>
</object>
    <object>
    <name>MCPFirst</name>
    <pose>Unspecified</pose>
    <truncated>0</truncated>
    <difficult>0</difficult>
    <bndbox>
        <xmin>827</xmin>
        <ymin>1100</ymin>
        <xmax>975</xmax>
        <ymax>1263</ymax>
    </bndbox>
</object>
    <object>
    <name>Ulna</name>
    <pose>Unspecified</pose>
    <truncated>0</truncated>
    <difficult>0</difficult>
    <bndbox>
        <xmin>513</xmin>
        <ymin>1334</ymin>
        <xmax>650</xmax>
        <ymax>1530</ymax>
    </bndbox>
</object>
    <object>
    <name>Radius</name>
    <pose>Unspecified</pose>
    <truncated>0</truncated>
    <difficult>0</difficult>
    <bndbox>
        <xmin>636</xmin>
        <ymin>1341</ymin>
        <xmax>832</xmax>
        <ymax>1567</ymax>
    </bndbox>
</object>
</annotation>

这里考虑到数据集整体比较规整，选择的是最为轻量级的n系列的模型，如下：

# YOLOv5 🚀 by Ultralytics, GPL-3.0 license

# Parameters
nc: 7  # number of classes
depth_multiple: 0.33  # model depth multiple
width_multiple: 0.25  # layer channel multiple
anchors:
  - [10,13, 16,30, 33,23]  # P3/8
  - [30,61, 62,45, 59,119]  # P4/16
  - [116,90, 156,198, 373,326]  # P5/32

# YOLOv5 v6.0 backbone
backbone:
  # [from, number, module, args]
  [[-1, 1, Conv, [64, 6, 2, 2]],  # 0-P1/2
   [-1, 1, Conv, [128, 3, 2]],  # 1-P2/4
   [-1, 3, C3, [128]],
   [-1, 1, Conv, [256, 3, 2]],  # 3-P3/8
   [-1, 6, C3, [256]],
   [-1, 1, Conv, [512, 3, 2]],  # 5-P4/16
   [-1, 9, C3, [512]],
   [-1, 1, Conv, [1024, 3, 2]],  # 7-P5/32
   [-1, 3, C3, [1024]],
   [-1, 1, SPPF, [1024, 5]],  # 9
  ]

# YOLOv5 v6.0 head
head:
  [[-1, 1, Conv, [512, 1, 1]],
   [-1, 1, nn.Upsample, [None, 2, 'nearest']],
   [[-1, 6], 1, Concat, [1]],  # cat backbone P4
   [-1, 3, C3, [512, False]],  # 13

   [-1, 1, Conv, [256, 1, 1]],
   [-1, 1, nn.Upsample, [None, 2, 'nearest']],
   [[-1, 4], 1, Concat, [1]],  # cat backbone P3
   [-1, 3, C3, [256, False]],  # 17 (P3/8-small)

   [-1, 1, Conv, [256, 3, 2]],
   [[-1, 14], 1, Concat, [1]],  # cat head P4
   [-1, 3, C3, [512, False]],  # 20 (P4/16-medium)

   [-1, 1, Conv, [512, 3, 2]],
   [[-1, 10], 1, Concat, [1]],  # cat head P5
   [-1, 3, C3, [1024, False]],  # 23 (P5/32-large)

   [[17, 20, 23], 1, Detect, [nc, anchors]],  # Detect(P3, P4, P5)
  ]

训练日志输出如下：