text-det-metric 0.0.2

Tool of computing the metric of text detection

Text Detect Metric

Evaluate on the custom dataset.

• Here we use the evaluation code of ch_mobile_v2_det on the text detection test set liekkas/text_det_test_dataset, and you can use the same analogy.

1. Install packages.

pip install modelscope==1.5.2
pip install text_det_metric

2. Run

1. Run get_pred_txt.py to get pred.txt

   import cv2
   from modelscope.msdatasets import MsDataset
   
   from ch_mobile_v2_det import TextDetector
   
   test_data = MsDataset.load(
       "text_det_test_dataset",
       namespace="liekkas",
       subset_name="default",
       split="test",
   )
   
   text_detector = TextDetector()
   
   content = []
   for one_data in test_data:
       img_path = one_data.get("image:FILE")
   
       print(img_path)
       img = cv2.imread(str(img_path))
       dt_boxes, scores, _ = text_detector(img)
       content.append(f"{img_path}\t{dt_boxes.tolist()}\t{scores}")
   
   with open("pred.txt", "w", encoding="utf-8") as f:
       for v in content:
           f.write(f"{v}\n")
   

2. Run compute_metric.py to get the metrics on the dataset

   from text_det_metric import DetectionIoUEvaluator
   
   metric = DetectionIoUEvaluator()
   
   # pred_path
   pred_path = "1.txt"
   mertric = metric(pred_path)
   print(mertric)
   

See details for TextDetMetric.