模型检测的结果，怎么返回 key_cls 参数？

[lemonit-eric-mao]

背景说明

使用标注数据的工具标注的 Label.txt 有个参数 key_cls 是用来记录分类的，如图

数据如下：

小图片/小02.png [{"transcription": "Table 2", "points": [[1, 2], [35, 2], [35, 14], [1, 14]], "difficult": false, "key_cls": "标题1"}, {"transcription": "Traralgon Ck", "points": [[2, 65], [74, 65], [74, 75], [2, 75]], "difficult": false, "key_cls": "名称"}, {"transcription": "Redhill", "points": [[2, 77], [74, 77], [74, 87], [2, 87]], "difficult": false, "key_cls": "名称"}, {"transcription": "Pine Ck", "points": [[2, 89], [74, 89], [74, 99], [2, 99]], "difficult": false, "key_cls": "名称"}, {"transcription": "Stewarts Ck 5", "points": [[2, 101], [74, 101], [74, 111], [2, 111]], "difficult": false, "key_cls": "名称"}, {"transcription": "Glendhu 2", "points": [[2, 113], [74, 113], [74, 123], [2, 123]], "difficult": false, "key_cls": "名称"}, {"transcription": "Cathedral Peak 2", "points": [[2, 125], [74, 125], [74, 135], [2, 135]], "difficult": false, "key_cls": "名称"}, {"transcription": "Cathedral Peak 3", "points": [[2, 135], [74, 135], [74, 145], [2, 145]], "difficult": false, "key_cls": "名称"}, {"transcription": "Lambrechtsbos A", "points": [[2, 148], [74, 148], [74, 158], [2, 158]], "difficult": false, "key_cls": "名称"}, {"transcription": "Lambrechtsbos B", "points": [[2, 159], [74, 159], [74, 169], [2, 169]], "difficult": false, "key_cls": "名称"}, {"transcription": "Biesievlei", "points": [[2, 171], [74, 171], [74, 181], [2, 181]], "difficult": false, "key_cls": "名称"}]

---

问题

示例代码

from paddleocr import PaddleOCR

ocr = PaddleOCR(
    use_doc_orientation_classify=False, # 通过 use_doc_orientation_classify 参数指定不使用文档方向分类模型
    use_doc_unwarping=False, # 通过 use_doc_unwarping 参数指定不使用文本图像矫正模型
    use_textline_orientation=False, # 通过 use_textline_orientation 参数指定不使用文本行方向分类模型
)
result = ocr.predict("./general_ocr_002.png")
for res in result:
    res.print()
    res.save_to_img("output")
    res.save_to_json("output")

那么使用 PP-OCRv5_server_det 进行检测时， 要如何也把 key_cls 参数给返回来？要怎么操？

[lemonit-eric-mao]