int
w2xconv_load_models(W2XConv *conv, const char *model_dir)
{
struct W2XConvImpl *impl = conv->impl;
std::string modelFileName(model_dir);
impl->noise1_models.clear();
impl->noise2_models.clear();
impl->scale2_models.clear();
if (!w2xc::modelUtility::generateModelFromJSON(modelFileName + "/noise1_model.json", impl->noise1_models)) {
setPathError(conv,
W2XCONV_ERROR_MODEL_LOAD_FAILED,
modelFileName + "/noise1_model.json");
return -1;
}
if (!w2xc::modelUtility::generateModelFromJSON(modelFileName + "/noise2_model.json", impl->noise2_models)) {
setPathError(conv,
W2XCONV_ERROR_MODEL_LOAD_FAILED,
modelFileName + "/noise2_model.json");
return -1;
}
if (!w2xc::modelUtility::generateModelFromJSON(modelFileName + "/scale2.0x_model.json", impl->scale2_models)) {
setPathError(conv,
W2XCONV_ERROR_MODEL_LOAD_FAILED,
modelFileName + "/scale2.0x_model.json");
return -1;
}
return 0;
}
void Waifu2xVideoFilter::Initialize(IScriptEnvironment* env) {
if (vi.pixel_type != VideoInfo::CS_YV12 &&
vi.pixel_type != VideoInfo::CS_YV16 &&
vi.pixel_type != VideoInfo::CS_YV24) {
env->ThrowError("Currently only YV12, YV16 or YV24 are supported.");
return;
}
if (nrLevel > 0) {
OutputDebugStringA("Waifu2x Loading NR Models.");
std::string modelFileName(modelsDir);
modelFileName = modelFileName + "\\noise" + std::to_string(this->nrLevel) + "_model.json";
if (!w2xc::modelUtility::generateModelFromJSON(modelFileName, this->modelsNR)) {
env->ThrowError("A noise model file is not found: %s", modelFileName.c_str());
return;
}
for (auto&& model : this->modelsNR) {
model->setNumberOfJobs(jobs);
}
}
if (enableScaling) {
OutputDebugStringA("Waifu2x Loading Scale Models.");
std::string modelFileName(modelsDir);
modelFileName = modelFileName + "/scale2.0x_model.json";
if (!w2xc::modelUtility::generateModelFromJSON(modelFileName, this->modelsScale)) {
env->ThrowError("A scale model file is not found: %s", modelFileName.c_str());
return;
}
for (auto&& model : this->modelsScale) {
model->setNumberOfJobs(jobs);
}
}
OutputDebugStringA("Waifu2x Finished Initializing.");
}
int main(int argc, char** argv) {
// definition of command line arguments
TCLAP::CmdLine cmd("waifu2x reimplementation using OpenCV", ' ', "1.0.0");
TCLAP::ValueArg<std::string> cmdInputFile("i", "input_file",
"path to input image file (you should input full path)", true, "",
"string", cmd);
TCLAP::ValueArg<std::string> cmdOutputFile("o", "output_file",
"path to output image file (you should input full path)", false,
"(auto)", "string", cmd);
std::vector<std::string> cmdModeConstraintV;
cmdModeConstraintV.push_back("noise");
cmdModeConstraintV.push_back("scale");
cmdModeConstraintV.push_back("noise_scale");
TCLAP::ValuesConstraint<std::string> cmdModeConstraint(cmdModeConstraintV);
TCLAP::ValueArg<std::string> cmdMode("m", "mode", "image processing mode",
false, "noise_scale", &cmdModeConstraint, cmd);
std::vector<int> cmdNRLConstraintV;
cmdNRLConstraintV.push_back(1);
cmdNRLConstraintV.push_back(2);
TCLAP::ValuesConstraint<int> cmdNRLConstraint(cmdNRLConstraintV);
TCLAP::ValueArg<int> cmdNRLevel("", "noise_level", "noise reduction level",
false, 1, &cmdNRLConstraint, cmd);
TCLAP::ValueArg<double> cmdScaleRatio("", "scale_ratio",
"custom scale ratio", false, 2.0, "double", cmd);
TCLAP::ValueArg<std::string> cmdModelPath("", "model_dir",
"path to custom model directory (don't append last / )", false,
"models", "string", cmd);
TCLAP::ValueArg<int> cmdNumberOfJobs("j", "jobs",
"number of threads launching at the same time", false, 4, "integer",
cmd);
// definition of command line argument : end
// parse command line arguments
try {
cmd.parse(argc, argv);
} catch (std::exception &e) {
std::cerr << e.what() << std::endl;
std::cerr << "Error : cmd.parse() threw exception" << std::endl;
std::exit(-1);
}
// load image file
cv::Mat image = cv::imread(cmdInputFile.getValue(), cv::IMREAD_COLOR);
image.convertTo(image, CV_32F, 1.0 / 255.0);
// set number of jobs for processing models
w2xc::modelUtility::getInstance().setNumberOfJobs(cmdNumberOfJobs.getValue());
// ===== Noise Reduction Phase =====
if (cmdMode.getValue() == "noise" || cmdMode.getValue() == "noise_scale") {
std::string modelFileName(cmdModelPath.getValue());
modelFileName = modelFileName + "/noise"
+ std::to_string(cmdNRLevel.getValue()) + "_model.json";
std::vector<std::unique_ptr<w2xc::Model> > models;
if (!w2xc::modelUtility::generateModelFromJSON(modelFileName, models))
std::exit(-1);
cv::Mat imageYUV;
cv::cvtColor(image, imageYUV, cv::COLOR_RGB2YUV);
std::vector<cv::Mat> imageSplit;
cv::Mat imageY;
cv::split(imageYUV, imageSplit);
imageSplit[0].copyTo(imageY);
w2xc::convertWithModels(imageY, imageSplit[0], models);
cv::merge(imageSplit, imageYUV);
cv::cvtColor(imageYUV, image, cv::COLOR_YUV2RGB);
} // noise reduction phase : end
// ===== scaling phase =====
if (cmdMode.getValue() == "scale" || cmdMode.getValue() == "noise_scale") {
// calculate iteration times of 2x scaling and shrink ratio which will use at last
int iterTimesTwiceScaling = static_cast<int>(std::ceil(
std::log2(cmdScaleRatio.getValue())));
double shrinkRatio = 0.0;
if (static_cast<int>(cmdScaleRatio.getValue())
!= std::pow(2, iterTimesTwiceScaling)) {
shrinkRatio = cmdScaleRatio.getValue()
/ std::pow(2.0, static_cast<double>(iterTimesTwiceScaling));
}
std::string modelFileName(cmdModelPath.getValue());
modelFileName = modelFileName + "/scale2.0x_model.json";
std::vector<std::unique_ptr<w2xc::Model> > models;
if (!w2xc::modelUtility::generateModelFromJSON(modelFileName, models))
std::exit(-1);
std::cout << "start scaling" << std::endl;
// 2x scaling
for (int nIteration = 0; nIteration < iterTimesTwiceScaling;
nIteration++) {
std::cout << "#" << std::to_string(nIteration + 1)
<< " 2x scaling..." << std::endl;
cv::Mat imageYUV;
cv::Size imageSize = image.size();
imageSize.width *= 2;
imageSize.height *= 2;
cv::Mat image2xNearest;
cv::resize(image, image2xNearest, imageSize, 0, 0, cv::INTER_NEAREST);
cv::cvtColor(image2xNearest, imageYUV, cv::COLOR_RGB2YUV);
std::vector<cv::Mat> imageSplit;
cv::Mat imageY;
cv::split(imageYUV, imageSplit);
imageSplit[0].copyTo(imageY);
// generate bicubic scaled image and
// convert RGB -> YUV and split
imageSplit.clear();
cv::Mat image2xBicubic;
cv::resize(image,image2xBicubic,imageSize,0,0,cv::INTER_CUBIC);
cv::cvtColor(image2xBicubic, imageYUV, cv::COLOR_RGB2YUV);
cv::split(imageYUV, imageSplit);
if(!w2xc::convertWithModels(imageY, imageSplit[0], models)){
std::cerr << "w2xc::convertWithModels : something error has occured.\n"
"stop." << std::endl;
std::exit(1);
};
cv::merge(imageSplit, imageYUV);
cv::cvtColor(imageYUV, image, cv::COLOR_YUV2RGB);
} // 2x scaling : end
if (shrinkRatio != 0.0) {
cv::Size lastImageSize = image.size();
lastImageSize.width =
static_cast<int>(static_cast<double>(lastImageSize.width
* shrinkRatio));
lastImageSize.height =
static_cast<int>(static_cast<double>(lastImageSize.height
* shrinkRatio));
cv::resize(image, image, lastImageSize, 0, 0, cv::INTER_LINEAR);
}
}
image.convertTo(image, CV_8U, 255.0);
std::string outputFileName = cmdOutputFile.getValue();
if (outputFileName == "(auto)") {
outputFileName = cmdInputFile.getValue();
int tailDot = outputFileName.find_last_of('.');
outputFileName.erase(tailDot, outputFileName.length());
outputFileName = outputFileName + "(" + cmdMode.getValue() + ")";
std::string &mode = cmdMode.getValue();
if(mode.find("noise") != mode.npos){
outputFileName = outputFileName + "(Level" + std::to_string(cmdNRLevel.getValue())
+ ")";
}
if(mode.find("scale") != mode.npos){
outputFileName = outputFileName + "(x" + std::to_string(cmdScaleRatio.getValue())
+ ")";
}
outputFileName += ".png";
}
cv::imwrite(outputFileName, image);
std::cout << "process successfully done!" << std::endl;
return 0;
}
