cyberagentailab / derendering-text Goto Github PK

I find that you use a func named 'sample_text_type_flags' in 'class OffsetSampler(object)' to define the angle. But it always return 0, maybe it is a error?

pip install -r requiements.txt
==> pip install -r requirements.txt

As you can see, some word images are out of the border, I think maybe there is a error in offset_sampler.py. I try to solve it, but there is still errors.

when I run this python gen.py --bgtype=color, there is a error:
[D 220704 06:56:21 handler:78] placement False [D 220704 06:56:21 gen_mode:105] 10, 0, 10000 Traceback (most recent call last): File "/home/yonghui_li/Image_generate/derendering-text/src/modules/generator/src/handler.py", line 32, in sample_and_render placed = self.sampler.sample(ih, dh) File "/home/yonghui_li/Image_generate/derendering-text/src/modules/generator/src/sampler.py", line 71, in sample style_data = self.style_sampler.sample(ih, dh) File "/home/yonghui_li/Image_generate/derendering-text/src/modules/generator/src/sampler_lib/style_sampler.py", line 30, in sample effect_params = self.get_style_params( File "/home/yonghui_li/Image_generate/derendering-text/src/modules/generator/src/sampler_lib/style_sampler.py", line 21, in get_style_params grad_param = skp.get_gradation_param( File "/home/yonghui_li/Image_generate/derendering-text/src/skia_lib/skia_paintor.py", line 85, in get_gradation_param color = get_color() NameError: name 'get_color' is not defined Traceback (most recent call last): File "/home/yonghui_li/Image_generate/derendering-text/src/modules/generator/src/handler.py", line 32, in sample_and_render placed = self.sampler.sample(ih, dh) File "/home/yonghui_li/Image_generate/derendering-text/src/modules/generator/src/sampler.py", line 71, in sample style_data = self.style_sampler.sample(ih, dh) File "/home/yonghui_li/Image_generate/derendering-text/src/modules/generator/src/sampler_lib/style_sampler.py", line 30, in sample effect_params = self.get_style_params( File "/home/yonghui_li/Image_generate/derendering-text/src/modules/generator/src/sampler_lib/style_sampler.py", line 21, in get_style_params grad_param = skp.get_gradation_param( File "/home/yonghui_li/Image_generate/derendering-text/src/skia_lib/skia_paintor.py", line 85, in get_gradation_param color = get_color() NameError: name 'get_color' is not defined

@shimoda-uec Hi, if I want to generate Chinese text images, what modifications do I need to make? Also, any advice to offer?

generate pre-rendered alpha maps for stroke
Traceback (most recent call last):
File "C:\ProgramData\Anaconda3\envs\derendering-text-master\lib\runpy.py", line 194, in _run_module_as_main
return _run_code(code, main_globals, None,
File "C:\ProgramData\Anaconda3\envs\derendering-text-master\lib\runpy.py", line 87, in _run_code
exec(code, run_globals)
File "C:\Users\mowe\Desktop\projects\derendering-text-master\util_lib\gen_pams.py", line 118, in
main()
File "C:\Users\mowe\Desktop\projects\derendering-text-master\util_lib\gen_pams.py", line 113, in main
gen_stroke_pams(char_dict)
File "C:\Users\mowe\Desktop\projects\derendering-text-master\util_lib\gen_pams.py", line 94, in gen_stroke_pams
alpha_crop = cv2.resize(
cv2.error: OpenCV(4.6.0) D:\a\opencv-python\opencv-python\opencv\modules\imgproc\src\resize.cpp:4052: error: (-215:Assertion failed) !ssize.empty() in function 'cv::resize'

Thank you for publishing the great work!

There are some mistakes in the bibtex in README.md.
The bibtex in the following link will be helpful.
https://openaccess.thecvf.com/content/ICCV2021/html/Shimoda_De-Rendering_Stylized_Texts_ICCV_2021_paper.html

Great Work! BTW, will the demo app be released? I would really like to try( it's so amzing😁

Hi, first of all let me congratulate you for this fantastic work.
After following the installation instructions I tried to run the usage example in cpu_mode
python test.py --imgfile=example/sample.jpg --use_cpu=1
and I get the following error:
[E 230209 19:16:52 test:125] Unexpected error has occurred
Traceback (most recent call last):
File "test.py", line 123, in
main(args)
File "test.py", line 88, in main
test(
File "test.py", line 53, in test
vd, rec_img = vectorize(pil_img, inps, outs)
File "/home/ariel/repos/derendering-text/src/modules/postprocess/vector.py", line 114, in vectorize
tb_param, effect_param, effect_visibility = extract_rendering_params(
File "/home/ariel/repos/derendering-text/src/modules/postprocess/manipulate.py", line 286, in extract_rendering_params
tb_param = get_textblob_param_with_affine(
File "/home/ariel/repos/derendering-text/src/modules/postprocess/rpe.py", line 515, in get_textblob_param_with_affine
text_form_dto = TextFormData(
TypeError: init() missing 1 required positional argument: 'text_index'

what I can see in the file dto_skia.py, is that the dataclass TextFromData requires a text_index: int value that is not set in rpe.py file. Please, let me know if you need more info in order to reproduce this error.
Best,
A.

I appreciate the provided code and tried myself.

For the prerendering script python -m util_lib.gen_pams

Following resize code causes error:

This is beause why alpha_crop = crop_alpha_area(alpha) results some (0,n) or (n,0) shaped array, therefore resize is impossible.
To deal with this, I temporarily altered code and checked how this issue as following:

and this is following logs:

Is this ok? Looks like lots of crop errors occur and result resizing error.