paddleOCR

Dockerfile

@@ -1,10 +1,11 @@
 FROM python:3.12.1
 
+ENV HOME /tmp
 ENV HF_HOME /tmp
 
 WORKDIR /code
 
-RUN apt-get update && apt-get install -y libgl1-mesa-glx libboost-thread1.74.0 && rm -rf /var/lib/apt/lists/*
+RUN apt-get update && apt-get install -y libgl1-mesa-glx libboost-thread1.74.0 libxext6 fonts-freefont-ttf && rm -rf /var/lib/apt/lists/*
 
 
 

main.py

@@ -6,11 +6,11 @@ from datetime import datetime
 from typing import List
 
 import cv2
-import easyocr
 import numpy as np
 from fastapi import FastAPI, HTTPException, Request, UploadFile
 from fastapi.middleware.cors import CORSMiddleware
 from fastapi.responses import JSONResponse
+from paddleocr import PaddleOCR
 from PIL import Image
 
 import dds
@@ -63,18 +63,10 @@ def load_ocr_model():
     グローバル変数readerに格納する。
     """
     global reader
-    model_dir = "/tmp/easyocr_models"
-    network_dir = "/tmp/easyocr_models/network"
 
     # 使用する言語と、モデルの保存先ディレクトリを指定してReaderを初期化
-    reader = easyocr.Reader(
-        ["en"],
-        model_storage_directory=model_dir,
-        user_network_directory=network_dir,
-    )
-    print(
-        f"EasyOCR model loaded successfully. Models are stored in {model_dir}"
-    )
+    reader = PaddleOCR(lang="en")
+    print(f"PaddleOCR model loaded successfully.")
 
 
 # @app.on_event("startup")
@@ -218,7 +210,7 @@ async def analyze_image(image_paths: list[UploadFile]):
                 text_mask = get_not_white_mask(img)
                 masked_img = cv2.bitwise_and(img, img, mask=text_mask)
                 # result = reader.readtext(candidate["img"])
-                result = reader.readtext(masked_img, text_threshold=0.3)
+                result = reader.ocr(masked_img)
                 print(result)
                 if len(result) > 0:
                     ocr_results.append(result)
@@ -240,7 +232,9 @@ async def analyze_image(image_paths: list[UploadFile]):
         # reds = []
         for i, result in enumerate(ocr_results):
             # text = result[0]["generated_text"].upper().strip()
-            _, text, _ = result[0]
+            text = result[0].get("rec_texts", [])
+            if text:
+                text = text[0]
             print(text, is_text_valid(text))
 
             text = is_text_valid(text)