Echo/fcb_photo_review
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

优化方向选择的条件

Browse Source
This commit is contained in:
liuyebo
2024-10-15 15:04:24 +08:00
parent 445d57e8c6
commit 947b4f20f3
1 changed files with 18 additions and 19 deletions
Download Patch File Download Diff File Expand all files Collapse all files

37
photo_review/auto_photo_review.py
View File

@@ -139,33 +139,32 @@ def information_extraction(phrec, pk_phhd):
else: # todo:也可能是图片，后续添加细分逻辑 else: # todo:也可能是图片，后续添加细分逻辑
dewarped_img_path = img_path dewarped_img_path = img_path
angles = model_util.clas_orientation(dewarped_img_path) angles = model_util.clas_orientation(dewarped_img_path)
ocr_result = []
ocr_text = '' ocr_text = ''
rotated_img = None info_extract = []
for angle in angles: for angle in angles:
tmp_ocr_result = [] ocr_result = []
tmp_rotated_img = image_util.rotate(dewarped_img_path, int(angle)) rotated_img = image_util.rotate(dewarped_img_path, int(angle))
split_results = image_util.split(tmp_rotated_img) split_results = image_util.split(rotated_img)
for split_result in split_results: for split_result in split_results:
if split_result['img'] is None: if split_result['img'] is None:
continue continue
a4_img = image_util.expand_to_a4_size(split_result['img']) a4_img = image_util.expand_to_a4_size(split_result['img'])
tmp_ocr_result += model_util.ocr(a4_img) ocr_result += model_util.ocr(a4_img)
tmp_ocr_text = common_util.ocr_result_to_text(tmp_ocr_result) tmp_ocr_text = common_util.ocr_result_to_text(ocr_result)
if len(tmp_ocr_text) > len(ocr_text): rec_type = model_util.clas_text(tmp_ocr_text) if ocr_text else None
ocr_result = tmp_ocr_result if rec_type == '基本医保结算单':
tmp_info_extract = model_util.ie_settlement(rotated_img, common_util.ocr_result_to_layout(ocr_result))
elif rec_type == '出院记录':
tmp_info_extract = model_util.ie_discharge(rotated_img, common_util.ocr_result_to_layout(ocr_result))
elif rec_type == '费用清单':
tmp_info_extract = model_util.ie_cost(rotated_img, common_util.ocr_result_to_layout(ocr_result))
else:
tmp_info_extract = None
if len(tmp_info_extract) > len(info_extract):
info_extract = tmp_info_extract
ocr_text = tmp_ocr_text ocr_text = tmp_ocr_text
rotated_img = tmp_rotated_img
rec_type = model_util.clas_text(ocr_text) if ocr_text else None
if rec_type == '基本医保结算单':
info_extract = model_util.ie_settlement(rotated_img, common_util.ocr_result_to_layout(ocr_result))
elif rec_type == '出院记录':
info_extract = model_util.ie_discharge(rotated_img, common_util.ocr_result_to_layout(ocr_result))
elif rec_type == '费用清单':
info_extract = model_util.ie_cost(rotated_img, common_util.ocr_result_to_layout(ocr_result))
else:
info_extract = None
if info_extract: if info_extract:
result_json = json.dumps(info_extract, ensure_ascii=False) result_json = json.dumps(info_extract, ensure_ascii=False)