Skip to content

πŸ“„ Awesome OCR multiple programing languages toolkits based on ONNX Runtime, OpenVINO, MNN, PaddlePaddle, TensorRT and PyTorch.

Add VisRes class

SWHL508beba09af5
diff --git a/python/demo.py b/python/demo.py
index 1cccf48..b791ce9 100644
--- a/python/demo.py
+++ b/python/demo.py
@@ -1,110 +1,24 @@
# -*- encoding: utf-8 -*-
# @Author: SWHL
# @Contact: [email protected]
-import math
-import random
-from pathlib import Path
-
import cv2
-import numpy as np
-from PIL import Image, ImageDraw, ImageFont
-
-from rapidocr_onnxruntime import RapidOCR
-
-# from rapidocr_openvino import RapidOCR
-
-
-def draw_ocr_box_txt(image, boxes, txts, font_path, scores=None, text_score=0.5):
- if not Path(font_path).exists():
- raise FileNotFoundError(
- f"The {font_path} does not exists! \n"
- f"Please download the file in the https://drive.google.com/file/d/1evWVX38EFNwTq_n5gTFgnlv8tdaNcyIA/view?usp=sharing"
- )
-
- h, w = image.height, image.width
- if image.mode == "L":
- image = image.convert("RGB")
-
- img_left = image.copy()
- img_right = Image.new("RGB", (w, h), (255, 255, 255))
-
- random.seed(0)
- draw_left = ImageDraw.Draw(img_left)
- draw_right = ImageDraw.Draw(img_right)
- for idx, (box, txt) in enumerate(zip(boxes, txts)):
- if scores is not None and float(scores[idx]) < text_score:
- continue
-
- color = (random.randint(0, 255), random.randint(0, 255), random.randint(0, 255))
- draw_left.polygon(box, fill=color)
- draw_right.polygon(
- [
- box[0][0],
- box[0][1],
- box[1][0],
- box[1][1],
- box[2][0],
- box[2][1],
- box[3][0],
- box[3][1],
- ],
- outline=color,
- )
-
- box_height = math.sqrt(
- (box[0][0] - box[3][0]) ** 2 + (box[0][1] - box[3][1]) ** 2
- )
-
- box_width = math.sqrt(
- (box[0][0] - box[1][0]) ** 2 + (box[0][1] - box[1][1]) ** 2
- )
-
- if box_height > 2 * box_width:
- font_size = max(int(box_width * 0.9), 10)
- font = ImageFont.truetype(font_path, font_size, encoding="utf-8")
- cur_y = box[0][1]
- for c in txt:
- char_size = font.getsize(c)
- draw_right.text((box[0][0] + 3, cur_y), c, fill=(0, 0, 0), font=font)
- cur_y += char_size[1]
- else:
- font_size = max(int(box_height * 0.8), 10)
- font = ImageFont.truetype(font_path, font_size, encoding="utf-8")
- draw_right.text([box[0][0], box[0][1]], txt, fill=(0, 0, 0), font=font)
-
- img_left = Image.blend(image, img_left, 0.5)
- img_show = Image.new("RGB", (w * 2, h), (255, 255, 255))
- img_show.paste(img_left, (0, 0, w, h))
- img_show.paste(img_right, (w, 0, w * 2, h))
- return np.array(img_show)
-
-
-def visualize(image_path, result, font_path="resources/fonts/FZYTK.TTF"):
- image = Image.open(image_path)
- boxes, txts, scores = list(zip(*result))
- draw_img = draw_ocr_box_txt(
- image, np.array(boxes), txts, font_path, scores, text_score=0.5
- )
+from rapidocr_onnxruntime import RapidOCR, VisRes
- draw_img_save = Path("./inference_results/")
- if not draw_img_save.exists():
- draw_img_save.mkdir(parents=True, exist_ok=True)
+# from rapidocr_openvino import RapidOCR, VisRes
- image_save = str(draw_img_save / f"infer_{Path(image_path).name}")
- cv2.imwrite(image_save, draw_img[:, :, ::-1])
- print(f"The infer result has saved in {image_save}")
+rapid_ocr = RapidOCR()
+vis = VisRes(font_path="resources/fonts/FZYTK.TTF")
-if __name__ == "__main__":
- rapid_ocr = RapidOCR()
+image_path = "tests/test_files/ch_en_num.jpg"
+with open(image_path, "rb") as f:
+ img = f.read()
- image_path = "tests/test_files/ch_en_num.jpg"
- with open(image_path, "rb") as f:
- img = f.read()
- result, elapse_list = rapid_ocr(img)
- print(result)
- print(elapse_list)
+result, elapse_list = rapid_ocr(img)
+print(result)
+print(elapse_list)
- if result:
- visualize(image_path, result, font_path="resources/fonts/FZYTK.TTF")
+boxes, txts, scores = list(zip(*result))
+res = vis(img, boxes, txts, scores)
+cv2.imwrite("vis.png", res)
diff --git a/python/rapidocr_onnxruntime/__init__.py b/python/rapidocr_onnxruntime/__init__.py
index ecc8206..c978620 100644
--- a/python/rapidocr_onnxruntime/__init__.py
+++ b/python/rapidocr_onnxruntime/__init__.py
@@ -2,4 +2,4 @@
# @Author: SWHL
# @Contact: [email protected]
from .main import RapidOCR
-from .utils import LoadImageError
+from .utils import LoadImageError, VisRes
diff --git a/python/rapidocr_onnxruntime/utils.py b/python/rapidocr_onnxruntime/utils.py
index 481c6c9..f4452ce 100644
--- a/python/rapidocr_onnxruntime/utils.py
+++ b/python/rapidocr_onnxruntime/utils.py
@@ -2,11 +2,13 @@
# @Author: SWHL
# @Contact: [email protected]
import argparse
+import math
+import random
import traceback
import warnings
from io import BytesIO
from pathlib import Path
-from typing import Dict, List, Optional, Union
+from typing import Dict, List, Optional, Tuple, Union
import cv2
import numpy as np
@@ -18,7 +20,7 @@ from onnxruntime import (
get_available_providers,
get_device,
)
-from PIL import Image, UnidentifiedImageError
+from PIL import Image, ImageDraw, ImageFont, UnidentifiedImageError
root_dir = Path(__file__).resolve().parent
InputType = Union[str, np.ndarray, bytes, Path]
@@ -344,3 +346,110 @@ class UpdateParameters:
k = k.split(prefix)[1]
new_rec_dict[k] = v
return new_rec_dict
+
+
+class VisRes:
+ def __init__(
+ self, font_path: Optional[Union[str, Path]] = None, text_score: float = 0.5
+ ):
+ if font_path is None:
+ raise FileNotFoundError(
+ f"The {font_path} does not exists! \n"
+ f"You could download the file in the https://drive.google.com/file/d/1evWVX38EFNwTq_n5gTFgnlv8tdaNcyIA/view?usp=sharing"
+ )
+
+ self.font_path = str(font_path)
+ self.text_score = text_score
+ self.load_img = LoadImage()
+
+ def __call__(
+ self,
+ img_content: InputType,
+ dt_boxes: np.ndarray,
+ txts: Optional[Union[List[str], Tuple[str]]] = None,
+ scores: Optional[Tuple[float]] = None,
+ ) -> np.ndarray:
+ img = self.load_img(img_content)
+ img = Image.fromarray(img)
+
+ if txts is None and scores is None:
+ return self.draw_dt_boxes(img, dt_boxes)
+
+ return self.draw_ocr_box_txt(img, dt_boxes, txts, scores)
+
+ def draw_dt_boxes(self, img: Image, dt_boxes: np.ndarray) -> np.ndarray:
+ img_temp = img.copy()
+ draw_img = ImageDraw.Draw(img_temp)
+ for idx, box in enumerate(dt_boxes):
+ draw_img.polygon(np.array(box), fill=self.get_random_color())
+
+ box_height = self.get_box_height(box)
+ font_size = max(int(box_height * 0.8), 10)
+ font = ImageFont.truetype(self.font_path, font_size, encoding="utf-8")
+ draw_img.polygon(
+ np.array(box).reshape(8).tolist(),
+ outline=(0, 0, 0),
+ )
+ draw_img.text([box[0][0], box[0][1]], str(idx), fill=(0, 0, 0), font=font)
+ return np.array(img_temp)
+
+ def draw_ocr_box_txt(self, image: Image, boxes, txts, scores=None):
+ h, w = image.height, image.width
+ if image.mode == "L":
+ image = image.convert("RGB")
+
+ img_left = image.copy()
+ img_right = Image.new("RGB", (w, h), (255, 255, 255))
+
+ random.seed(0)
+ draw_left = ImageDraw.Draw(img_left)
+ draw_right = ImageDraw.Draw(img_right)
+ for idx, (box, txt) in enumerate(zip(boxes, txts)):
+ if scores is not None and float(scores[idx]) < self.text_score:
+ continue
+
+ color = self.get_random_color()
+ draw_left.polygon(np.array(box), fill=color)
+ draw_right.polygon(
+ np.array(box).reshape(8).tolist(),
+ outline=color,
+ )
+
+ box_height = self.get_box_height(box)
+ box_width = self.get_box_width(box)
+ if box_height > 2 * box_width:
+ font_size = max(int(box_width * 0.9), 10)
+ font = ImageFont.truetype(self.font_path, font_size, encoding="utf-8")
+ cur_y = box[0][1]
+ for c in txt:
+ char_size = font.getsize(c)
+ draw_right.text(
+ (box[0][0] + 3, cur_y), c, fill=(0, 0, 0), font=font
+ )
+ cur_y += char_size[1]
+ else:
+ font_size = max(int(box_height * 0.8), 10)
+ font = ImageFont.truetype(self.font_path, font_size, encoding="utf-8")
+ draw_right.text([box[0][0], box[0][1]], txt, fill=(0, 0, 0), font=font)
+
+ img_left = Image.blend(image, img_left, 0.5)
+ img_show = Image.new("RGB", (w * 2, h), (255, 255, 255))
+ img_show.paste(img_left, (0, 0, w, h))
+ img_show.paste(img_right, (w, 0, w * 2, h))
+ return np.array(img_show)
+
+ @staticmethod
+ def get_random_color():
+ return (
+ random.randint(0, 255),
+ random.randint(0, 255),
+ random.randint(0, 255),
+ )
+
+ @staticmethod
+ def get_box_height(box: List[List[float]]) -> float:
+ return math.sqrt((box[0][0] - box[3][0]) ** 2 + (box[0][1] - box[3][1]) ** 2)
+
+ @staticmethod
+ def get_box_width(box: List[List[float]]) -> float:
+ return math.sqrt((box[0][0] - box[1][0]) ** 2 + (box[0][1] - box[1][1]) ** 2)
diff --git a/python/rapidocr_openvino/__init__.py b/python/rapidocr_openvino/__init__.py
index ecc8206..c978620 100644
--- a/python/rapidocr_openvino/__init__.py
+++ b/python/rapidocr_openvino/__init__.py
@@ -2,4 +2,4 @@
# @Author: SWHL
# @Contact: [email protected]
from .main import RapidOCR
-from .utils import LoadImageError
+from .utils import LoadImageError, VisRes
diff --git a/python/rapidocr_openvino/utils.py b/python/rapidocr_openvino/utils.py
index 093e487..32aadbb 100644
--- a/python/rapidocr_openvino/utils.py
+++ b/python/rapidocr_openvino/utils.py
@@ -2,15 +2,17 @@
# @Author: SWHL
# @Contact: [email protected]
import argparse
+import math
+import random
from io import BytesIO
from pathlib import Path
-from typing import Dict, List, Optional, Union
+from typing import Dict, List, Optional, Tuple, Union
import cv2
import numpy as np
import yaml
from openvino.runtime import Core
-from PIL import Image, UnidentifiedImageError
+from PIL import Image, ImageDraw, ImageFont, UnidentifiedImageError
root_dir = Path(__file__).resolve().parent
InputType = Union[str, np.ndarray, bytes, Path]
@@ -273,3 +275,110 @@ class UpdateParameters:
k = k.split(prefix)[1]
new_rec_dict[k] = v
return new_rec_dict
+
+
+class VisRes:
+ def __init__(
+ self, font_path: Optional[Union[str, Path]] = None, text_score: float = 0.5
+ ):
+ if font_path is None:
+ raise FileNotFoundError(
+ f"The {font_path} does not exists! \n"
+ f"You could download the file in the https://drive.google.com/file/d/1evWVX38EFNwTq_n5gTFgnlv8tdaNcyIA/view?usp=sharing"
+ )
+
+ self.font_path = str(font_path)
+ self.text_score = text_score
+ self.load_img = LoadImage()
+
+ def __call__(
+ self,
+ img_content: InputType,
+ dt_boxes: np.ndarray,
+ txts: Optional[Union[List[str], Tuple[str]]] = None,
+ scores: Optional[Tuple[float]] = None,
+ ) -> np.ndarray:
+ img = self.load_img(img_content)
+ img = Image.fromarray(img)
+
+ if txts is None and scores is None:
+ return self.draw_dt_boxes(img, dt_boxes)
+
+ return self.draw_ocr_box_txt(img, dt_boxes, txts, scores)
+
+ def draw_dt_boxes(self, img: Image, dt_boxes: np.ndarray) -> np.ndarray:
+ img_temp = img.copy()
+ draw_img = ImageDraw.Draw(img_temp)
+ for idx, box in enumerate(dt_boxes):
+ draw_img.polygon(np.array(box), fill=self.get_random_color())
+
+ box_height = self.get_box_height(box)
+ font_size = max(int(box_height * 0.8), 10)
+ font = ImageFont.truetype(self.font_path, font_size, encoding="utf-8")
+ draw_img.polygon(
+ np.array(box).reshape(8).tolist(),
+ outline=(0, 0, 0),
+ )
+ draw_img.text([box[0][0], box[0][1]], str(idx), fill=(0, 0, 0), font=font)
+ return np.array(img_temp)
+
+ def draw_ocr_box_txt(self, image: Image, boxes, txts, scores=None):
+ h, w = image.height, image.width
+ if image.mode == "L":
+ image = image.convert("RGB")
+
+ img_left = image.copy()
+ img_right = Image.new("RGB", (w, h), (255, 255, 255))
+
+ random.seed(0)
+ draw_left = ImageDraw.Draw(img_left)
+ draw_right = ImageDraw.Draw(img_right)
+ for idx, (box, txt) in enumerate(zip(boxes, txts)):
+ if scores is not None and float(scores[idx]) < self.text_score:
+ continue
+
+ color = self.get_random_color()
+ draw_left.polygon(np.array(box), fill=color)
+ draw_right.polygon(
+ np.array(box).reshape(8).tolist(),
+ outline=color,
+ )
+
+ box_height = self.get_box_height(box)
+ box_width = self.get_box_width(box)
+ if box_height > 2 * box_width:
+ font_size = max(int(box_width * 0.9), 10)
+ font = ImageFont.truetype(self.font_path, font_size, encoding="utf-8")
+ cur_y = box[0][1]
+ for c in txt:
+ char_size = font.getsize(c)
+ draw_right.text(
+ (box[0][0] + 3, cur_y), c, fill=(0, 0, 0), font=font
+ )
+ cur_y += char_size[1]
+ else:
+ font_size = max(int(box_height * 0.8), 10)
+ font = ImageFont.truetype(self.font_path, font_size, encoding="utf-8")
+ draw_right.text([box[0][0], box[0][1]], txt, fill=(0, 0, 0), font=font)
+
+ img_left = Image.blend(image, img_left, 0.5)
+ img_show = Image.new("RGB", (w * 2, h), (255, 255, 255))
+ img_show.paste(img_left, (0, 0, w, h))
+ img_show.paste(img_right, (w, 0, w * 2, h))
+ return np.array(img_show)
+
+ @staticmethod
+ def get_random_color():
+ return (
+ random.randint(0, 255),
+ random.randint(0, 255),
+ random.randint(0, 255),
+ )
+
+ @staticmethod
+ def get_box_height(box: List[List[float]]) -> float:
+ return math.sqrt((box[0][0] - box[3][0]) ** 2 + (box[0][1] - box[3][1]) ** 2)
+
+ @staticmethod
+ def get_box_width(box: List[List[float]]) -> float:
+ return math.sqrt((box[0][0] - box[1][0]) ** 2 + (box[0][1] - box[1][1]) ** 2)