统计数据集中各个类别实例数量

import json
import pandas as pd
from collections import defaultdict

def count_class_instances(json_file):
    class_counts = defaultdict(int)
    with open(json_file, 'r') as file:
        data = json.load(file)
        annotations = data['annotations']
    for annotation in annotations:
        category_id = annotation['category_id']
        class_counts[category_id] += 1
    return class_counts

# 文件路径
train_json_file = 'path/to/train.json'
val_json_file = 'path/to/val.json'
test_json_file = 'path/to/test.json'

# 统计
train_class_counts = count_class_instances(train_json_file)
val_class_counts = count_class_instances(val_json_file)
test_class_counts = count_class_instances(test_json_file)

# 创建DataFrame
data = {
    'Train': pd.Series(train_class_counts),
    'Validation': pd.Series(val_class_counts),
    'Test': pd.Series(test_class_counts)
}
df = pd.DataFrame(data).fillna(0)  # 使用0填充NaN值
df.index.name = 'Category ID'  # 设置索引名称
df = df.astype(int)  # 确保所有计数都是整数

# 添加总计行
df.loc['Total'] = df.sum()

# 打印表格
print(df)