Comments (9)
ku21fan
commented on May 13, 2024
5
Hi, we have test.py for inference,
and currently, I am going to make demo.py for the simplified version of test.py.
I will upload demo.py in this weak.
Thanks.
from deep-text-recognition-benchmark.
ku21fan
commented on May 13, 2024
5
Sorry for the late update.
I just uploaded demo.py.
check the run demo part in the Readme.
Best regards.
from deep-text-recognition-benchmark.
sharavsambuu
commented on May 13, 2024
4
Hello, great repo!
Here is my naive attempt.
https://colab.research.google.com/drive/1PHnc_QYyf9b1_KJ1r15wYXaOXkdm1Mrk
from deep-text-recognition-benchmark.
sharavsambuu
commented on May 13, 2024
1
Hi @hasiftwistcode if you want to use best_accuracy.pth, please take an attention on the parser.character, it is using 36 characters '0123456789abcdefghijklmnopqrstuvwxyz'. Other one(case-sensitive version) is using 96 characters, string.printable[:-6]. I have not evaluated an accuracy of my dataset yet.
from deep-text-recognition-benchmark.
hszkf
commented on May 13, 2024
1
The model can even predict my input image though, but it is alright. Next time, ill use this model to train first on my dataset. Thanks for your help!
from deep-text-recognition-benchmark.
sharavsambuu
commented on May 13, 2024
1
I updated notebook for your 4 images. It seems result is not good.
from deep-text-recognition-benchmark.
hszkf
commented on May 13, 2024
Thanks @sharavsambuu ! I changed it to best_accuracy.pth and it didn't work, is there any reason why?
And the model is pretty inaccurate with any type of images right? like this : https://drive.google.com/drive/folders/15SMmcdY9fzeKg6aHZjqlsfbqWNn2i6_0?usp=sharing
from deep-text-recognition-benchmark.
hszkf
commented on May 13, 2024
Appreciate your effort! I've looked at it and yeah as expected. I am currently doing ocr for the identity card, I did crop the address to little pieces which the images are really small and surprisingly tesseract is pretty good at it (very image-sensitive). I tried kraken ocr but it wasn't good as tesseract.
from deep-text-recognition-benchmark.
sharavsambuu
commented on May 13, 2024
Thank you for letting me know, I also working on similar project. Tried various repos and results were not so good, also tried google vision api and it is working perfect except their high api price. Also tried tesseract and it needs a lot of preprocessing in order to work well in my case. Currently scraping and collecting my own dataset in order to cope those problems :) I hope you will do well, cheers!
from deep-text-recognition-benchmark.
Related Issues (20)
- Poor performance đ after training on custom data âšī¸ HOT 1
- Can I run train.py ignoring characters that are in the dataset but not in the --character? HOT 1
- When ground truth is incorrect but, prediction is well-predicted. HOT 3
- Training stuck after first iteration
- rtmp://live.restream.io/live
- training success ,but failed is run as python file
- training success , but run python code error, about
- Text Color Recognition in Images
- after Fine tuning, demo.py can't decode original data HOT 1
- RuntimeError HOT 1
- how to get coordinate of the text
- how i convert .zip files to .pth?
- AssertionError: datasets should not be an empty iterable
- CUDA error, while training Attn as prediction
- Cannot train model for upper case characters
- I encountered a bug when executing the following command in the deep-text-recognition-benchmark repository HOT 1
- [Question] Configuring the dataset for custom training HOT 2
- custom dataset Calculated padded input size per channel: (1 x 22). Kernel size: (2 x 2)
- Check the wrong prediction data
- Can it be used to extract texts from an entire page ?
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
đ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. đđđ
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google â¤ī¸ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from deep-text-recognition-benchmark.