Ask Your Question
0

How to use Text Detection Advanced_EAST with OpenCV DNN

asked 2019-03-26 05:29:16 -0600

Hi, Advanced_EAST is an algorithm used for Scene image text detect, which is primarily based on EAST, and the significant improvement was also made, which make long text predictions more accurate . I want to call Advanced_EAST with OpenCV DNN just like EAST. I tried many ways but I failed. Can you help me? You can find Advanced_EAST here: https://github.com/huoyijie/AdvancedEAST , and my .pb model is here: https://drive.google.com/open?id=1CQ7.... Thanks!

edit retag flag offensive close merge delete

Comments

Have you solved it?

louzhu gravatar imagelouzhu ( 2019-08-27 06:20:41 -0600 )edit

1 answer

Sort by » oldest newest most voted
0

answered 2019-03-27 04:21:50 -0600

dkurt gravatar image

updated 2019-03-27 04:22:16 -0600

Please add more details about

I tried many ways but I failed. Can you help me?

What kind of problem you have? I tried to run this mode using the latest state of OpenCV and it runs fine:

import cv2 as cv
import numpy as np

net = cv.dnn.readNet('east_model_3T256.pb')
inp = np.random.standard_normal([1, 3, 320, 320]).astype(np.float32)
net.setInput(inp)
out = net.forward()
print(out.shape)  # (1, 7, 80, 80)
edit flag offensive delete link more

Comments

Hi,

Thank you for your replay. I do't know how to get "inside_score"、"side_vertex_code" and "side_vertex_coord" separately from the "out" and how to use "inside_score"、"side_vertex_code" and "side_vertex_coord" to detect text from an image.Can you give me a more detailed demo to show me how to use this .pb file to detect text from an image.

jack.he gravatar imagejack.he ( 2019-03-27 04:47:38 -0600 )edit

@jack.he, please read the documentation or use help(net) in python. There is a method which produces multiple outputs.

dkurt gravatar imagedkurt ( 2019-03-27 05:23:54 -0600 )edit

Question Tools

1 follower

Stats

Asked: 2019-03-26 05:29:16 -0600

Seen: 1,187 times

Last updated: Mar 27 '19