• Stars
    star
    459
  • Rank 95,377 (Top 2 %)
  • Language
    Python
  • Created almost 5 years ago
  • Updated almost 5 years ago

Reviews

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

Repository Details

Generate your lover with your photo

CoupleGenerator

Introduction

Generate your lover with your photo

This project is an old project from two years ago. I accidentally found it in my storage medium.

We collect pictures of married couples on the Internet and pre-process the images.

The model was trained by using the code of pix2pix.

An example of training images:

image

You can make photos of yourself with the person you loved as a training pair~

This repo is just for entertainment.

If you have any interests (not for business), please feel free to use the code and the dataset.

Fitting the training data

Here are the results of the training set after 8800 steps. image

The model can fit the training images in a short time.

Environments

Tensorflow==1.1

You also need to download the pretrain weight for vgg through following links:

https://github.com/machrisaa/tensorflow-vgg

Quick start

Download the training images and unzip it: https://cloudstor.aarnet.edu.au/plus/s/VWZJaWfbla3kFch

Run bash autotest.sh

Trained model (with limited data and training steps): https://cloudstor.aarnet.edu.au/plus/s/YHDWgez1g3RFc6o

We use all the data for training and testing, it is not scientific for research. The motivation of this project is a surprise for my friend. We want to overfit her with the boy she loves secretly. Therefore we do not consider the generalization of the model.

Extension

The marriage images were collected on BaiDu image by using a spider. Here is a very old and simple code for collecting images: https://blog.csdn.net/qq_27879381/article/details/65015280#comments

Just change the keyword.

Thanks GZHermit for helping me to process the data. He detected the face from the collecting images and re-located the couple into training pairs.

The images we provide is just for fun, not for business activities

If you really want to train an effective model for business, you have to collect more images in legal way.

AD

This project is a toy project, and I am working on semantic segmentation, instance segmentation on 2D images and videos for my Ph.D. career. If you are interested in these topics, you can follow my github. I will release a new code on video segmentation recently.

CVPR2019: https://github.com/irfanICMLL/structure_knowledge_distillation

More Repositories

1

structure_knowledge_distillation

The official code for the paper 'Structured Knowledge Distillation for Semantic Segmentation'. (CVPR 2019 ORAL) and extension to other tasks.
Python
699
star
2

ETC-Real-time-Per-frame-Semantic-video-segmentation

Enforcing temporal consistency in real-time per-frame semantic video segmentation
Python
296
star
3

TorchDistiller

Python
192
star
4

Auto_painter

Recently, realistic image generation using deep neural networks has become a hot topic in machine learning and computer vision. Such an image can be generated at pixel level by learning from a large collection of images. Learning to generate colorful cartoon images from black-and-white sketches is not only an interesting research problem, but also a useful application in digital entertainment. In this paper, we investigate the sketch-to-image synthesis problem by using conditional generative adversarial networks (cGAN). We propose a model called auto-painter which can automatically generate compatible colors given a sketch. Wasserstein distance is used in training cGAN to overcome model collapse and enable the model converged much better. The new model is not only capable of painting hand-draw sketch with compatible colors, but also allowing users to indicate preferred colors. Experimental results on different sketch datasets show that the auto-painter performs better than other existing image-to-image methods.
Python
132
star
5

EMM-for-stock-prediction

We propose a model to analyze sentiment of online stock forum and use the information to predict stock volatility in the Chinese market. By generating a sentimental dictionary, we analyze the sentimental tendencies of each post as sentiment indicators. Such sentimental information will be fused with market data for prediction based on Recurrent Neural Networks (RNNs). We manually labeled the sentiment of forum post and make the data public available for research. Empirical evidence shows that 8 of the 10 stocks perform better with sentimental indicators.
Python
62
star
6

Auto_painter_demo

The code of building a web demo for Auto_painter
JavaScript
27
star
7

SSIW

The code of 'The devil is in the labels: Semantic segmentation from sentences'.
Python
13
star
8

inceptionV2_finetune

Fine-tuning of inceptionV2 on CUB-200 Birds dataset in Tensorflow
Python
9
star
9

stock_predict

This project predicts stock trends on the basis of online user comments and LSTM
Python
5
star
10

colorization

reading note
3
star
11

horseSeg

raw_code
Python
1
star
12

dvn

dvn for semantic segmentation
1
star