YOLO相同等级关键点标注修改方案

项介绍：

使用YOLOv8关键点检测之前需要进行标注，在人体标注的时候由于每个关键点都是不一样的，在标注时需要进行区分各个关键点。但是如果标注大量相同等级的关键点时，如果进行分别标注效率会变低，因此考虑进行统一标注后，再修改其标签，具体如下：

---

问题描述

在标注具有相同等级的关键点时，无法对每个点进行区分，且一个一个选标签耽误时间。

人体的关键点都是不相同的
而对于章鱼脚来说，各个脚的都是相同等级的关键点！!

如果想要进行区分标注十分困难，因此考虑统一标注相同的关键点，后续进行修改！！！！

1.首选统一标注如下：

2.得到标注JSON文件

{
  "version": "5.1.1",
  "flags": {},
  "shapes": [
    {
      "label": "foot",
      "points": [
        [
          176.1094674556213,
          45.95857988165682
        ]
      ],
      "group_id": null,
      "shape_type": "point",
      "flags": {}
    },
    {
      "label": "foot",
      "points": [
        [
          335.28106508875743,
          42.70414201183433
        ]
      ],
      "group_id": null,
      "shape_type": "point",
      "flags": {}
    },
    {
      "label": "foot",
      "points": [
        [
          400.9615384615385,
          216.66863905325448
        ]
      ],
      "group_id": null,
      "shape_type": "point",
      "flags": {}
    },
    {
      "label": "foot",
      "points": [
        [
          264.2751479289941,
          324.9526627218935
        ]
      ],
      "group_id": null,
      "shape_type": "point",
      "flags": {}
    },
    {
      "label": "foot",
      "points": [
        [
          117.52958579881656,
          208.97633136094677
        ]
      ],
      "group_id": null,
      "shape_type": "point",
      "flags": {}
    }
  ],

3.运行修改代码

import os
import json

# 载入JSON数据
def load_json(json_path):
    with open(json_path, 'r') as file:
        data = json.load(file)
    return data

# 判断点是否在矩形框内
def is_point_in_rect(point, rect):
    x_min = min(rect[0][0], rect[1][0])
    x_max = max(rect[0][0], rect[1][0])
    y_min = min(rect[0][1], rect[1][1])
    y_max = max(rect[0][1], rect[1][1])
    return x_min <= point[0] <= x_max and y_min <= point[1] <= y_max

# 处理JSON数据，寻找并重命名关键点
def process_data(data):
    for shape in data['shapes']:
        if shape['shape_type'] == 'rectangle':
            # 找到矩形框
            rect = shape['points']
            # 寻找矩形内的所有关键点
            seed_heads = [s for s in data['shapes'] if s['shape_type'] == 'point' and s['label'] == 'Seed_head' and is_point_in_rect(s['points'][0], rect)]
            # 对关键点进行排序并重新命名
            seed_heads.sort(key=lambda x: (x['points'][0][0], x['points'][0][1]))  # 按x坐标和y坐标排序
            for idx, head in enumerate(seed_heads):
                head['label'] = f'foot{idx + 1}'

# 主函数
def main():
    input_folder = 'keypoint'  # 输入文件夹路径
    output_folder = 'keypoint1'  # 输出文件夹路径

    # 确保输出文件夹存在，若不存在则创建
    if not os.path.exists(output_folder):
        os.makedirs(output_folder)

    # 遍历输入文件夹中的所有文件
    for filename in os.listdir(input_folder):
        if filename.endswith('.json'):
            json_path = os.path.join(input_folder, filename)
            output_path = os.path.join(output_folder, filename)
            data = load_json(json_path)
            process_data(data)
            # 保存修改后的数据
            with open(output_path, 'w') as f:
                json.dump(data, f, indent=4)

main()

需要修改的位置！！！！！！！！！

seed_heads = [s for s in data['shapes'] if s['shape_type'] == 'point' and s['label'] == 'foot' and is_point_in_rect(s['points'][0], rect)]#将foot修改成你的标签

head['label'] = f'foot{idx + 1}'#将foot修改成你的标签

input_folder = 'keypoint'  # 输入文件夹路径
output_folder = 'keypoint1'  # 输出文件夹路径

结果：

下面是修改后的JSON文件

{
    "version": "5.1.1",
    "flags": {},
    "shapes": [
        {
            "label": "foot2",
            "points": [
                [
                    176.1094674556213,
                    45.95857988165682
                ]
            ],
            "group_id": null,
            "shape_type": "point",
            "flags": {}
        },
        {
            "label": "foot4",
            "points": [
                [
                    335.28106508875743,
                    42.70414201183433
                ]
            ],
            "group_id": null,
            "shape_type": "point",
            "flags": {}
        },
        {
            "label": "foot5",
            "points": [
                [
                    400.9615384615385,
                    216.66863905325448
                ]
            ],
            "group_id": null,
            "shape_type": "point",
            "flags": {}
        },
        {
            "label": "foot3",
            "points": [
                [
                    264.2751479289941,
                    324.9526627218935
                ]
            ],
            "group_id": null,
            "shape_type": "point",
            "flags": {}
        },
        {
            "label": "foot1",
            "points": [
                [
                    117.52958579881656,
                    208.97633136094677
                ]
            ],
            "group_id": null,
            "shape_type": "point",
            "flags": {}
        },
        {
            "label": "star",
            "points": [
                [
                    102.22868217054264,
                    30.984496124031068
                ],
                [
                    414.63178294573646,
                    339.5116279069768
                ]
            ],
            "group_id": null,
            "shape_type": "rectangle",
            "flags": {}
        }
    ],

标签修改成功

总结

1.标注时可以选用相同名称进行标注节省时间
2.使用代码对JSON文件进行修改

1.找到框
2.找到所有点
3.判断哪个点在哪个框里后修改

3.完成标注关键点