Miss detections on Small Size text with same font

[sharoseali]

Hi, thanks for your great work. I am applying PPOCR on a pdf page having architectural drawings. The problem is some text is not been detected, however same font text in the same page is also detected at a smaller distance. For example, in the below image some of the C3 and C4 text is detected but some are missed. Can anyone help with this? How can I resolve this issue?
The detector I am using is DB++

Thanks

[UserWangZz]

1. Try the ppocr-v4 version of the model
2. During inference, try to change det_limit_side_len, det_db_box_thresh
   det_db_score_mode and other parameters

You can refer to the following code：

PaddleOCR/tools/infer/utility.py

Lines 53 to 68 in 4336771

	# params for text detector
	parser.add_argument("--image_dir", type=str)
	parser.add_argument("--page_num", type=int, default=0)
	parser.add_argument("--det_algorithm", type=str, default="DB")
	parser.add_argument("--det_model_dir", type=str)
	parser.add_argument("--det_limit_side_len", type=float, default=960)
	parser.add_argument("--det_limit_type", type=str, default="max")
	parser.add_argument("--det_box_type", type=str, default="quad")
	# DB parmas
	parser.add_argument("--det_db_thresh", type=float, default=0.3)
	parser.add_argument("--det_db_box_thresh", type=float, default=0.6)
	parser.add_argument("--det_db_unclip_ratio", type=float, default=1.5)
	parser.add_argument("--max_batch_size", type=int, default=10)
	parser.add_argument("--use_dilation", type=str2bool, default=False)
	parser.add_argument("--det_db_score_mode", type=str, default="fast")

[sharoseali]

Thanks for your reply. I have updated the my ppocr python package via pip to 2.6.1. The downloaded model's folder starts from ch like ch_PP-OCRv4_server_det. How can I download the English language model? Among all the models shown below, which detection and recognition models are the the best in performance while inference.

thanks