File size: 3,670 Bytes
9633a4f 5533d5d 9633a4f 5533d5d |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 |
import os
import torch
print(torch.__version__)
# torch_ver, cuda_ver = torch.__version__.split('+')
# os.system('pip list')
# os.system(f'pip install opencv-contrib-python==4.5.5.62 --no-cache-dir')
# os.system('pip list')
# os.system(f'pip install pycocotools==2.0.0 mmdet mmcv-full==1.5.0 -f https://download.openmmlab.com/mmcv/dist/{cuda_ver}/torch1.10.0/index.html --no-cache-dir')
# os.system('wget -nv -c https://download.openmmlab.com/mmocr/data/wildreceipt.tar; mkdir -p data; tar -xf wildreceipt.tar --directory data; rm -f wildreceipt.tar')
import datetime
import gradio as gr
import pandas as pd
from mmocr.utils.ocr import MMOCR
import os
def inference(img, det, recog):
print(datetime.datetime.now(), 'start')
# ocr = MMOCR(det='PS_CTW', recog='SAR', kie='SDMGR')
# ocr = MMOCR(det=det, recog=recog, kie='SDMGR')
if det == 'None':
det = None
if recog == 'None':
recog = None
ocr = MMOCR(det=det, recog=recog)
print(datetime.datetime.now(), 'start read:', img.name)
results = ocr.readtext(img.name, details=True, output='output')
result_file = 'output/out_{}.png'.format(os.path.splitext(os.path.basename(img.name))[0])
print(datetime.datetime.now(), results)
# return result_file, pd.DataFrame(results[0]['result']).iloc[: , 2:]
return result_file, results
description = 'Gradio demo for MMOCR. MMOCR is an open-source toolbox based on PyTorch and mmdetection for text detection, text recognition, and the corresponding downstream tasks including key information extraction. To use it, simply upload your image or click one of the examples to load them. Read more at the links below.'
article = "<p style='text-align: center'><a href='https://mmocr.readthedocs.io/en/latest/'>MMOCR is an open-source toolbox based on PyTorch and mmdetection for text detection, text recognition, and the corresponding downstream tasks including key information extraction.</a> | <a href='https://github.com/open-mmlab/mmocr'>Github Repo</a></p>"
examples = []
path = './images'
files = os.listdir(path)
files.sort()
for f in files:
file = os.path.join(path, f)
if os.path.isfile(file):
examples.append([file, 'PS_CTW', 'SAR'])
det = gr.inputs.Dropdown(choices=[
'DB_r18',
'DB_r50',
'DBPP_r50',
'DRRG',
'FCE_IC15',
'FCE_CTW_DCNv2',
'MaskRCNN_CTW',
'MaskRCNN_IC15',
'MaskRCNN_IC17',
'PANet_CTW',
'PANet_IC15',
'PS_CTW',
'PS_IC15',
'TextSnake',
'None'
], type="value", default='PS_CTW', label='det')
recog = gr.inputs.Dropdown(choices=[
'CRNN',
'SAR',
'SAR_CN',
'NRTR_1/16-1/8',
'NRTR_1/8-1/4',
'RobustScanner',
'SATRN',
'SATRN_sm',
'ABINet',
'ABINet_Vision',
'SEG',
'CRNN_TPS',
'MASTER',
'None'
], type="value", default='SAR', label='recog')
gr.Interface(inference,
[gr.inputs.Image(type='file', label='Input'), det, recog ],
# [gr.outputs.Image(type='pil', label='Output'), gr.outputs.Dataframe(headers=['text', 'text_score', 'label', 'label_score'])],
[gr.outputs.Image(type='pil', label='Output'), gr.outputs.Textbox(type='str', label='Prediction')],
title='MMOCR',
description=description,
article=article,
examples=examples,
css=".output_image, .input_image {height: 40rem !important; width: 100% !important;}",
enable_queue=True
).launch(debug=True, server_name='0.0.0.0', server_port=7860)
|