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

优化图片标识

Browse Source
This commit is contained in:
liuyebo
2024-06-28 16:12:55 +08:00
parent 3645e13f00
commit 7e4b17c30e
1 changed files with 19 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files

30
photo_review/photo_review.py
View File

@@ -9,7 +9,6 @@ import urllib.request
import cv2 import cv2
import numpy as np import numpy as np
import paddle
import paddleclas import paddleclas
from paddlenlp import Taskflow from paddlenlp import Taskflow
@@ -18,7 +17,8 @@ sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
from time import sleep from time import sleep
from sqlalchemy import update from sqlalchemy import update
from config.keys import PATIENT_NAME, ADMISSION_DATE, DISCHARGE_DATE, MEDICAL_EXPENSES, PERSONAL_CASH_PAYMENT, \ from config.keys import PATIENT_NAME, ADMISSION_DATE, DISCHARGE_DATE, MEDICAL_EXPENSES, PERSONAL_CASH_PAYMENT, \
PERSONAL_ACCOUNT_PAYMENT, PERSONAL_FUNDED_AMOUNT, MEDICAL_INSURANCE_TYPE, HOSPITAL, DEPARTMENT, DOCTOR, SETTLEMENT_LIST_SCHEMA, DISCHARGE_RECORD_SCHEMA, COST_LIST_SCHEMA PERSONAL_ACCOUNT_PAYMENT, PERSONAL_FUNDED_AMOUNT, MEDICAL_INSURANCE_TYPE, HOSPITAL, DEPARTMENT, DOCTOR, \
SETTLEMENT_LIST_SCHEMA, DISCHARGE_RECORD_SCHEMA, COST_LIST_SCHEMA
from config.mysql import MysqlSession from config.mysql import MysqlSession
from config.photo_review import PHHD_BATCH_SIZE, SLEEP_MINUTES, LAYOUT_ANALYSIS from config.photo_review import PHHD_BATCH_SIZE, SLEEP_MINUTES, LAYOUT_ANALYSIS
from photo_review.entity.bd_yljg import BdYljg from photo_review.entity.bd_yljg import BdYljg
@@ -74,7 +74,8 @@ def split_image(img_path, max_ratio=2.82, best_ration=1.41, overlap=0.05):
split_result.append({"img": cropped_img, "x_offset": 0, "y_offset": offset}) split_result.append({"img": cropped_img, "x_offset": 0, "y_offset": offset})
# 最后一次裁剪时不足的部分填充黑色 # 最后一次裁剪时不足的部分填充黑色
last_img = split_result[-1]["img"] last_img = split_result[-1]["img"]
split_result[-1]["img"] = cv2.copyMakeBorder(last_img, 0, round(cropped_width - last_img.shape[0]), 0, 0, cv2.BORDER_CONSTANT, value=(0, 0, 0)) split_result[-1]["img"] = cv2.copyMakeBorder(last_img, 0, round(cropped_width - last_img.shape[0]), 0, 0,
cv2.BORDER_CONSTANT, value=(0, 0, 0))
else: else:
# 宽度是较长边 # 宽度是较长边
cropped_height = height * best_ration cropped_height = height * best_ration
@@ -84,7 +85,8 @@ def split_image(img_path, max_ratio=2.82, best_ration=1.41, overlap=0.05):
split_result.append({"img": cropped_img, "x_offset": offset, "y_offset": 0}) split_result.append({"img": cropped_img, "x_offset": offset, "y_offset": 0})
# 最后一次裁剪时不足的部分填充黑色 # 最后一次裁剪时不足的部分填充黑色
last_img = split_result[-1]["img"] last_img = split_result[-1]["img"]
split_result[-1]["img"] = cv2.copyMakeBorder(last_img, 0, 0, 0, round(cropped_height - last_img.shape[1]), cv2.BORDER_CONSTANT, value=(0, 0, 0)) split_result[-1]["img"] = cv2.copyMakeBorder(last_img, 0, 0, 0, round(cropped_height - last_img.shape[1]),
cv2.BORDER_CONSTANT, value=(0, 0, 0))
else: else:
split_result.append({"img": img, "x_offset": 0, "y_offset": 0}) split_result.append({"img": img, "x_offset": 0, "y_offset": 0})
return split_result return split_result
@@ -134,14 +136,14 @@ def rotate_image(img, angle):
# 关键信息提取 # 关键信息提取
def information_extraction(ie, phrecs): def information_extraction(ie, phrecs):
result = {} result = {}
# 同一批图的标识
identity = int(time.time())
for phrec in phrecs: for phrec in phrecs:
pic_path = ucloud.get_private_url(phrec.cfjaddress) pic_path = ucloud.get_private_url(phrec.cfjaddress)
if not pic_path: if not pic_path:
continue continue
split_result = split_image(pic_path) split_result = split_image(pic_path)
# 同一张图的标识
identity = int(time.time())
for img in split_result: for img in split_result:
with tempfile.NamedTemporaryFile(delete=False, suffix=".jpg") as temp_file: with tempfile.NamedTemporaryFile(delete=False, suffix=".jpg") as temp_file:
angle = get_image_rotation_angle(img["img"]) angle = get_image_rotation_angle(img["img"])
@@ -165,17 +167,14 @@ def information_extraction(ie, phrecs):
result_json = result_json[:5000] result_json = result_json[:5000]
session = MysqlSession() session = MysqlSession()
zx_ocr = ZxOcr(pk_phhd=phrec.pk_phhd, pk_phrec=phrec.pk_phrec, id=identity, cfjaddress=phrec.cfjaddress, zx_ocr = ZxOcr(pk_phhd=phrec.pk_phhd, pk_phrec=phrec.pk_phrec, id=identity, cfjaddress=phrec.cfjaddress,
content=result_json, rotation_angle=angle, x_offset=img["x_offset"], y_offset=img["y_offset"], content=result_json, rotation_angle=angle, x_offset=img["x_offset"],
create_time=now, update_time=now) y_offset=img["y_offset"], create_time=now, update_time=now)
session.add(zx_ocr) session.add(zx_ocr)
session.commit() session.commit()
session.close() session.close()
result = merge_result(result, ie_result) result = merge_result(result, ie_result)
# # 完成一次任务释放显存
# paddle.device.cuda.empty_cache()
# logging.info("显存释放完成")
return result return result
@@ -249,7 +248,8 @@ def photo_review(pk_phhd):
elif phrec.cRectype == "4": elif phrec.cRectype == "4":
cost_list.append(phrec) cost_list.append(phrec)
settlement_list_ie_result = information_extraction(Taskflow("information_extraction", schema=SETTLEMENT_LIST_SCHEMA, model="uie-x-base", settlement_list_ie_result = information_extraction(
Taskflow("information_extraction", schema=SETTLEMENT_LIST_SCHEMA, model="uie-x-base",
task_path="config/model/settlement_list_model", layout_analysis=LAYOUT_ANALYSIS), settlement_list) task_path="config/model/settlement_list_model", layout_analysis=LAYOUT_ANALYSIS), settlement_list)
settlement_data = { settlement_data = {
"pk_phhd": pk_phhd, "pk_phhd": pk_phhd,
@@ -276,7 +276,8 @@ def photo_review(pk_phhd):
settlement_data["personal_funded_amount"] = handle_decimal(settlement_data["personal_funded_amount_str"]) settlement_data["personal_funded_amount"] = handle_decimal(settlement_data["personal_funded_amount_str"])
save_or_update_ie(ZxIeSettlement, pk_phhd, settlement_data) save_or_update_ie(ZxIeSettlement, pk_phhd, settlement_data)
discharge_record_ie_result = information_extraction(Taskflow("information_extraction", schema=DISCHARGE_RECORD_SCHEMA, model="uie-x-base", discharge_record_ie_result = information_extraction(
Taskflow("information_extraction", schema=DISCHARGE_RECORD_SCHEMA, model="uie-x-base",
task_path="config/model/discharge_record_model", layout_analysis=LAYOUT_ANALYSIS), discharge_record) task_path="config/model/discharge_record_model", layout_analysis=LAYOUT_ANALYSIS), discharge_record)
discharge_data = { discharge_data = {
"pk_phhd": pk_phhd, "pk_phhd": pk_phhd,
@@ -314,7 +315,8 @@ def photo_review(pk_phhd):
discharge_data["department"] = ylks.name discharge_data["department"] = ylks.name
save_or_update_ie(ZxIeDischarge, pk_phhd, discharge_data) save_or_update_ie(ZxIeDischarge, pk_phhd, discharge_data)
cost_list_ie_result = information_extraction(Taskflow("information_extraction", schema=COST_LIST_SCHEMA, model="uie-x-base", cost_list_ie_result = information_extraction(
Taskflow("information_extraction", schema=COST_LIST_SCHEMA, model="uie-x-base",
task_path="config/model/cost_list_model", layout_analysis=LAYOUT_ANALYSIS), cost_list) task_path="config/model/cost_list_model", layout_analysis=LAYOUT_ANALYSIS), cost_list)
cost_data = { cost_data = {
"pk_phhd": pk_phhd, "pk_phhd": pk_phhd,