b01902041 / deep-virtual-try-on-with-clothes-transform

An image-based virtual try-on system with deep learning.

Deep Virtual Try-on with Clothes Transform

Source code for paper "Deep Virtual Try-on with Clothes Transform"

Overall Architecture

Dependencies

Install dependencies using pip.

pip install -r requirements.txt

Step1: CAGAN

code and data

• Training: CAGAN.py

python CAGAN.py

• Testing: Testing_with_fixed_data.py

python Testing_with_fixed_data.py

• Data: MVC_image_pairs_resize_new.zip

parameters in code

Training: CAGAN.py

• Data should be put in

"./MVC_image_pairs_resize_new/1/*.jpg" (for person images)

"./MVC_image_pairs_resize_new/5/*.jpg" (for clothes images)

470: data = "data folder name"

471: train_A = "person images folder name"

473: filenames_1 = "person images folder name"

474: filenames_5 = "clothes images folder name"

617, 618: set "save model path"

Testing: Testing_with_fixed_data.py

• Data should be put in

"./MVC_image_pairs_resize_new/1/*.jpg" (for person images)

"./MVC_image_pairs_resize_new/5_test/*.jpg" (for clothes images)

215: set "model path"

220: data = "data folder name"

221: train_A = "person images folder name"

222: filenames_5 = "clothes images folder name"

224: out_root_dir = "output folder name"

225: origin_dir = "save input person images"

226: target_dir = "save target clothes images"

227: output_dir = "save output images"

228: mask_dir = "save output masks"

230: testing_number = "how much data you want to test"

Step2: Segmentation

code

https://github.com/Engineering-Course/LIP_SSL

• Modify mask: modify_mask.m

• Save the masks file to png file: show.m

• Combine all the masks: combine_with_CAGANmask.m

Step3: Transform

code and data

• Training: unet.py data.py

python unet.py

• Testing: Testing_unet.py

python Testing_unet.py

• Data: transform_data.zip transform_test_data.zip

parameters in code

Training: unet.py

336: model_dir = "save model path"

337: result_dir = "save results path"

223: set "loss type"

data.py

15: set "data path"

Testing: Testing_unet.py

16: test_data_path = "data path"

17: test_img_folder = "target clothes image folder name"

18: test_mask_folder = "mask folder name"

19: model_name = "model name"

20: result_dir = "save results path"

Step4: Combination

code

Combine_image.m

Results

deep-virtual-try-on-with-clothes-transform's Issues

How to run step2

Hi thanks for the code it is really helpful, I could use step 1, but could you help me with step 2, will the output generated from step 1 be enough for step 2? Or step 2 needs more other special inputs ? Also do the folder addresses in 'combine_with_CAGANmask.m' point to outputs of step 1?

training_updates = Adam(lr=lrD, beta_1=0.5).get_updates(weightsD,[],loss_D) TypeError: get_updates() takes 3 positional arguments but 4 were given"

I'm getting this error when I'm trying to run the "CAGAN.py" code.

I tried this code in python versions 3.7,3.8,3.9,3.10.
I'm using Tensorflow version 2.11.0

Please help me in resolving this issue.

It is a great work! Can you share your dataset?

It is a great work! Can you share your dataset?
My email [email protected]

requirements.txt

Could You upload requrement file.

Its being very hard for me to run your code.

Thanks in advance

whoever did this please check your gmail

No Output results in Step3

Blank black images in results after transformation step(Step3)

CAGAN.py is not running

Thanks for sharing the code. Great work! Hey, I just played around with this code using MVC data but the model CAGAN.py is not running. I guess because the data have not been loaded properly. Would you be able to share the data directory/structure? Thanks so much again.
Cheers,
Jen

How to run this project?

This code's result is very good. Thank you for your sharing. Could you teach me how to run this code?
Thank you very much.

What is the Python version?

Hi, thanks for sharing the code, can you tell me what the Python version is?

I can't find output mask M' with Step 1. CAGAN

Hi ~
Step1. CAGAN: I got result_512 only and I can't find mask M' ( no output in source code)

Could you help me to solve this code ?

Thank you very much!