Table Transformer做表格检测和识别实践

计算机视觉方面的三大顶级会议：ICCV,CVPR,ECCV.统称ICE
CVPR 2022文档图像分析与识别相关论文26篇汇集简介

论文： PubTables-1M: Towards comprehensive table extraction from unstructured documents是发表于CVPR上的一篇论文
作者发布了两个模型，表格检测和表格结构识别。

论文讲解可以参考【论文阅读】PubTables- 1M: Towards comprehensive table extraction from unstructured documents

hugging face Table Transformer 使用文档
hugging face Table DETR 使用文档

检测表格

from huggingface_hub import hf_hub_download
from transformers import AutoImageProcessor, TableTransformerForObjectDetection
import torch
from PIL import Imagefile_path = hf_hub_download(repo_id="nielsr/example-pdf", repo_type="dataset", filename="example_pdf.png")image = Image.open(file_path).convert("RGB")image_processor = AutoImageProcessor.from_pretrained("microsoft/table-transformer-detection")
model = TableTransformerForObjectDetection.from_pretrained("microsoft/table-transformer-detection")inputs = image_processor(images=image, return_tensors="pt")
outputs = model(**inputs)# convert outputs (bounding boxes and class logits) to COCO API
target_sizes = torch.tensor([image.size[::-1]])
results = image_processor.post_process_object_detection(outputs, threshold=0.9, target_sizes=target_sizes)[0
]for score, label, box in zip(results["scores"], results["labels"], results["boxes"]):box = [round(i, 2) for i in box.tolist()]print(f"Detected {model.config.id2label[label.item()]} with confidence "f"{round(score.item(), 3)} at location {box}")region = image.crop(box) #检测region.save('xxx.jpg') #保存# Detected table with confidence 1.0 at location [202.1, 210.59, 1119.22, 385.09]

结果 ：效果不错

表格结构识别

参考：https://github.com/NielsRogge/Transformers-Tutorials/blob/master/Table%20Transformer/Using_Table_Transformer_for_table_detection_and_table_structure_recognition.ipynb

import torch
from PIL import Image
from transformers import DetrFeatureExtractor
from transformers import AutoImageProcessor, TableTransformerForObjectDetection
from huggingface_hub import hf_hub_downloadfeature_extractor = DetrFeatureExtractor()file_path = hf_hub_download(repo_id="nielsr/example-pdf", repo_type="dataset", filename="example_pdf.png")
image = Image.open(file_path).convert("RGB")encoding = feature_extractor(image, return_tensors="pt")
model = TableTransformerForObjectDetection.from_pretrained("microsoft/table-transformer-structure-recognition")with torch.no_grad():outputs = model(**encoding)
target_sizes = [image.size[::-1]]
results = feature_extractor.post_process_object_detection(outputs, threshold=0.6, target_sizes=target_sizes)[0]
# plot_results(image, results['scores'], results['labels'], results['boxes'])
results

获取列图像:

columns_box_list = [results['boxes'][i].tolist() for i in range(len(results['boxes'])) if results['labels'][i].item()==1] columns_1 = image.crop(columns_box_list[0]) 
columns_1.save('columns_1.jpg') #保存

可视化：

import matplotlib.pyplot as plt
# colors for visualization
COLORS = [[0.000, 0.447, 0.741], [0.850, 0.325, 0.098], [0.929, 0.694, 0.125],[0.494, 0.184, 0.556], [0.466, 0.674, 0.188], [0.301, 0.745, 0.933]]def plot_results(pil_img, scores, labels, boxes):plt.figure(figsize=(16, 10))plt.imshow(pil_img)ax = plt.gca()colors = COLORS * 100for score, label, (xmin, ymin, xmax, ymax), c in zip(scores.tolist(), labels.tolist(), boxes.tolist(), colors):ax.add_patch(plt.Rectangle((xmin, ymin), xmax - xmin, ymax - ymin,fill=False, color=c, linewidth=3))text = f'{model.config.id2label[label]}: {score:0.2f}'ax.text(xmin, ymin, text, fontsize=15,bbox=dict(facecolor='yellow', alpha=0.5))plt.axis('off')plt.show()

post_process_object_detection方法：

OpenCV PIL图像格式互转

参考：https://blog.csdn.net/dcrmg/article/details/78147219

PIL–》OpenCV

cv2.cvtColor(numpy.asarray(image),cv2.COLOR_RGB2BGR)

import cv2
from PIL import Image
import numpyimage = Image.open("plane.jpg")
image.show()
img = cv2.cvtColor(numpy.asarray(image),cv2.COLOR_RGB2BGR)
cv2.imshow("OpenCV",img)
cv2.waitKey()

OpenCV --》 PIL

 Image.fromarray(cv2.cvtColor(img,cv2.COLOR_BGR2RGB))

import cv2
from PIL import Image
import numpyimg = cv2.imread("plane.jpg")
cv2.imshow("OpenCV",img)
image = Image.fromarray(cv2.cvtColor(img,cv2.COLOR_BGR2RGB))
image.show()
cv2.waitKey()