mirror of
https://github.com/opencv/opencv.git
synced 2026-07-31 08:13:04 +04:00
Merge pull request #27051 from gursimarsingh:move_ccm_to_photo_module
Adding color correction module to photo module from opencv_contrib #27051 This PR moved color correction module from opencv_contrib to main repo inside photo module. ### Pull Request Readiness Checklist See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request - [x] I agree to contribute to the project under Apache 2 License. - [x] To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV - [x] The PR is proposed to the proper branch - [x] There is a reference to the original bug report and related work - [x] There is accuracy test, performance test and test data in opencv_extra repository, if applicable Patch to opencv_extra has the same branch name. - [x] The feature is well documented and sample code can be built with the project CMake
This commit is contained in:
Gursimar Singh
GitHub
@@ -0,0 +1,224 @@
|
||||
//! [tutorial]
|
||||
#include <opencv2/core.hpp>
|
||||
#include <opencv2/highgui.hpp>
|
||||
#include <opencv2/imgcodecs.hpp>
|
||||
#include <opencv2/imgproc.hpp>
|
||||
#include <opencv2/photo.hpp>
|
||||
#include <opencv2/objdetect.hpp>
|
||||
#include <opencv2/dnn.hpp>
|
||||
#include <iostream>
|
||||
#include "../dnn/common.hpp"
|
||||
|
||||
using namespace std;
|
||||
using namespace cv;
|
||||
using namespace cv::dnn;
|
||||
using namespace cv::ccm;
|
||||
using namespace mcc;
|
||||
|
||||
const string about =
|
||||
"This sample detects Macbeth color checker using DNN or thresholding and applies color correction."
|
||||
"To run default:\n"
|
||||
"\t ./example_cpp_color_correction_model --input=path/to/your/input/image --query=path/to/your/query/image\n"
|
||||
"With DNN model:\n"
|
||||
"\t ./example_cpp_color_correction_model mcc --input=path/to/your/input/image --query=path/to/your/query/image\n\n"
|
||||
"Using pre-computed CCM:\n"
|
||||
"\t ./example_cpp_color_correction_model mcc --ccm_file=path/to/ccm_output.yaml --query=path/to/your/query/image\n\n"
|
||||
"Model path can also be specified using --model argument. And config path can be specified using --config. Download it using python download_models.py mcc from dnn samples directory\n\n";
|
||||
|
||||
const string param_keys =
|
||||
"{ help h | | Print help message. }"
|
||||
"{ @alias | | An alias name of model to extract preprocessing parameters from models.yml file. }"
|
||||
"{ zoo | ../dnn/models.yml | An optional path to file with preprocessing parameters }"
|
||||
"{ input i | mcc_ccm_test.jpg | Path to input image for computing CCM.}"
|
||||
"{ query q | baboon.jpg | Path to query image to apply color correction. If not provided, input image will be used. }"
|
||||
"{ type | 0 | chartType: 0-Standard, 1-DigitalSG, 2-Vinyl }"
|
||||
"{ num_charts | 1 | Maximum number of charts in the image }"
|
||||
"{ ccm_file | | Path to YAML file containing pre-computed CCM parameters}";
|
||||
|
||||
const string backend_keys = format(
|
||||
"{ backend | default | Choose one of computation backends: "
|
||||
"default: automatically (by default), "
|
||||
"openvino: Intel's Deep Learning Inference Engine (https://software.intel.com/openvino-toolkit), "
|
||||
"opencv: OpenCV implementation, "
|
||||
"vkcom: VKCOM, "
|
||||
"cuda: CUDA, "
|
||||
"webnn: WebNN }");
|
||||
|
||||
const string target_keys = format(
|
||||
"{ target | cpu | Choose one of target computation devices: "
|
||||
"cpu: CPU target (by default), "
|
||||
"opencl: OpenCL, "
|
||||
"opencl_fp16: OpenCL fp16 (half-float precision), "
|
||||
"vpu: VPU, "
|
||||
"vulkan: Vulkan, "
|
||||
"cuda: CUDA, "
|
||||
"cuda_fp16: CUDA fp16 (half-float preprocess) }");
|
||||
|
||||
string keys = param_keys + backend_keys + target_keys;
|
||||
|
||||
static bool processFrame(const Mat& frame, Ptr<CCheckerDetector> detector, Mat& src, int nc){
|
||||
if (!detector->process(frame, nc))
|
||||
{
|
||||
return false;
|
||||
}
|
||||
vector<Ptr<CChecker>> checkers = detector->getListColorChecker();
|
||||
src = checkers[0]->getChartsRGB(false);
|
||||
|
||||
return true;
|
||||
}
|
||||
|
||||
int main(int argc, char* argv[]) {
|
||||
CommandLineParser parser(argc, argv, keys);
|
||||
parser.about(about);
|
||||
|
||||
if (parser.has("help")) {
|
||||
cout << about << endl;
|
||||
parser.printMessage();
|
||||
return 0;
|
||||
}
|
||||
|
||||
string modelName = parser.get<String>("@alias");
|
||||
string zooFile = parser.get<String>("zoo");
|
||||
const char* path = getenv("OPENCV_SAMPLES_DATA_PATH");
|
||||
|
||||
if ((path != NULL) || parser.has("@alias")) {
|
||||
zooFile = findFile(zooFile);
|
||||
}
|
||||
else{
|
||||
cout<<"[WARN] set the environment variables or pass the arguments --model, --config and models.yml file using --zoo for using dnn based detector. Continuing with default detector.\n\n";
|
||||
}
|
||||
keys += genPreprocArguments(modelName, zooFile);
|
||||
parser = CommandLineParser(argc, argv, keys);
|
||||
|
||||
int t = parser.get<int>("type");
|
||||
if (t < 0 || t > 2)
|
||||
{
|
||||
cout << "Error: --type must be 0, 1 or 2" << endl;
|
||||
parser.printMessage(); // prints full usage
|
||||
return -1;
|
||||
}
|
||||
|
||||
ColorChart chartType = ColorChart(t);
|
||||
|
||||
const string sha1 = parser.get<String>("sha1");
|
||||
const string modelPath = findModel(parser.get<string>("model"), sha1);
|
||||
const string config_sha1 = parser.get<String>("config_sha1");
|
||||
const string configPath = findModel(parser.get<string>("config"), config_sha1);
|
||||
const string backend = parser.get<String>("backend");
|
||||
const string target = parser.get<String>("target");
|
||||
|
||||
int nc = parser.get<int>("num_charts");
|
||||
|
||||
// Get input and target image paths
|
||||
const string inputFile = parser.get<String>("input");
|
||||
const string queryFile = parser.get<String>("query");
|
||||
const string ccmFile = parser.get<String>("ccm_file");
|
||||
|
||||
if (!ccmFile.empty()) {
|
||||
// When ccm_file is provided, only query is required
|
||||
if (queryFile.empty()) {
|
||||
cout << "Error: Query image path must be provided when using pre-computed CCM." << endl;
|
||||
parser.printMessage();
|
||||
return -1;
|
||||
}
|
||||
} else {
|
||||
// Original validation for when computing new CCM
|
||||
if (inputFile.empty()) {
|
||||
cout << "Error: Input image path must be provided." << endl;
|
||||
parser.printMessage();
|
||||
return -1;
|
||||
}
|
||||
}
|
||||
|
||||
ColorCorrectionModel model;
|
||||
Mat queryImage;
|
||||
|
||||
if (!ccmFile.empty()) {
|
||||
// Load CCM from YAML file
|
||||
FileStorage fs(ccmFile, FileStorage::READ);
|
||||
if (!fs.isOpened()) {
|
||||
cout << "Error: Unable to open CCM file: " << ccmFile << endl;
|
||||
return -1;
|
||||
}
|
||||
model.read(fs["ColorCorrectionModel"]);
|
||||
fs.release();
|
||||
cout << "Loaded CCM from file: " << ccmFile << endl;
|
||||
|
||||
// Read query image when using pre-computed CCM
|
||||
queryImage = imread(findFile(queryFile));
|
||||
if (queryImage.empty()) {
|
||||
cout << "Error: Unable to read query image." << endl;
|
||||
return -1;
|
||||
}
|
||||
} else {
|
||||
// Read input image for computing new CCM
|
||||
Mat originalImage = imread(findFile(inputFile));
|
||||
if (originalImage.empty()) {
|
||||
cout << "Error: Unable to read input image." << endl;
|
||||
return -1;
|
||||
}
|
||||
|
||||
// Process first image to compute CCM
|
||||
Mat image = originalImage.clone();
|
||||
Mat src;
|
||||
|
||||
Ptr<CCheckerDetector> detector;
|
||||
if (!modelPath.empty() && !configPath.empty()) {
|
||||
Net net = readNetFromTensorflow(modelPath, configPath);
|
||||
net.setPreferableBackend(getBackendID(backend));
|
||||
net.setPreferableTarget(getTargetID(target));
|
||||
detector = CCheckerDetector::create(net);
|
||||
cout << "Using DNN-based checker detector." << endl;
|
||||
} else {
|
||||
detector = CCheckerDetector::create();
|
||||
cout << "Using thresholding-based checker detector." << endl;
|
||||
}
|
||||
detector->setColorChartType(chartType);
|
||||
|
||||
if (!processFrame(image, detector, src, nc)) {
|
||||
cout << "No chart detected in the input image!" << endl;
|
||||
return -1;
|
||||
}
|
||||
// Convert to double and normalize
|
||||
src.convertTo(src, CV_64F, 1.0/255.0);
|
||||
|
||||
// Color correction model
|
||||
model = ColorCorrectionModel(src, COLORCHECKER_MACBETH);
|
||||
model.setCcmType(CCM_LINEAR);
|
||||
model.setDistance(DISTANCE_CIE2000);
|
||||
model.setLinearization(LINEARIZATION_GAMMA);
|
||||
model.setLinearizationGamma(2.2);
|
||||
|
||||
Mat ccm = model.compute();
|
||||
cout << "Computed CCM Matrix:\n" << ccm << endl;
|
||||
cout << "Loss: " << model.getLoss() << endl;
|
||||
|
||||
// Save model parameters to YAML file
|
||||
FileStorage fs("ccm_output.yaml", FileStorage::WRITE);
|
||||
model.write(fs);
|
||||
fs.release();
|
||||
cout << "Model parameters saved to ccm_output.yaml" << endl;
|
||||
|
||||
// Set query image for correction
|
||||
if (queryFile.empty()) {
|
||||
cout << "[WARN] No query image provided, applying color correction on input image" << endl;
|
||||
queryImage = originalImage.clone();
|
||||
} else {
|
||||
queryImage = imread(findFile(queryFile));
|
||||
if (queryImage.empty()) {
|
||||
cout << "Error: Unable to read query image." << endl;
|
||||
return -1;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
Mat calibratedImage;
|
||||
model.correctImage(queryImage, calibratedImage);
|
||||
|
||||
imshow("Original Image", queryImage);
|
||||
imshow("Corrected Image", calibratedImage);
|
||||
waitKey(0);
|
||||
|
||||
return 0;
|
||||
}
|
||||
//! [tutorial]
|
||||
@@ -0,0 +1,24 @@
|
||||
0.380463 0.31696 0.210053
|
||||
0.649781 0.520561 0.452553
|
||||
0.323114 0.37593 0.50123
|
||||
0.314785 0.396522 0.258116
|
||||
0.452971 0.418602 0.578767
|
||||
0.34908 0.608649 0.652283
|
||||
0.691127 0.517818 0.144984
|
||||
0.208668 0.224391 0.485851
|
||||
0.657849 0.378126 0.304115
|
||||
0.285762 0.229671 0.31913
|
||||
0.513422 0.685031 0.337381
|
||||
0.786459 0.676133 0.246303
|
||||
0.11751 0.135079 0.383441
|
||||
0.190745 0.470513 0.296844
|
||||
0.587832 0.299132 0.196117
|
||||
0.783908 0.746261 0.294357
|
||||
0.615481 0.359983 0.471403
|
||||
0.107095 0.370516 0.573142
|
||||
0.708598 0.718936 0.740915
|
||||
0.593812 0.612474 0.63222
|
||||
0.489774 0.510077 0.521757
|
||||
0.380591 0.398499 0.393662
|
||||
0.27461 0.293267 0.275244
|
||||
0.180753 0.194968 0.145006
|
||||
Binary file not shown.
|
After Width: | Height: | Size: 72 KiB |
@@ -0,0 +1,185 @@
|
||||
import cv2 as cv
|
||||
import numpy as np
|
||||
import argparse
|
||||
import sys
|
||||
import os
|
||||
sys.path.append(os.path.join(os.path.dirname(__file__), ".."))
|
||||
from dnn.common import *
|
||||
|
||||
def get_args_parser(func_args):
|
||||
backends = ("default", "openvino", "opencv", "vkcom", "cuda", "webnn")
|
||||
targets = ("cpu", "opencl", "opencl_fp16", "vpu", "vulkan", "cuda", "cuda_fp16")
|
||||
|
||||
parser = argparse.ArgumentParser(add_help=False)
|
||||
parser.add_argument('--zoo', default=os.path.join(os.path.dirname(os.path.abspath(__file__)), '../dnn', 'models.yml'),
|
||||
help='An optional path to file with preprocessing parameters.')
|
||||
parser.add_argument('--input', default='mcc_ccm_test.jpg', help='Path to input image for computing CCM')
|
||||
parser.add_argument('--query', default='baboon.jpg', help='Path to query image to apply color correction')
|
||||
parser.add_argument('--ccm_file', help='Path to YAML file containing pre-computed CCM parameters')
|
||||
parser.add_argument('--chart_type', type=int, default=0,
|
||||
help='chartType: 0-Standard, 1-DigitalSG, 2-Vinyl, default:0')
|
||||
parser.add_argument('--num_charts', type=int, default=1,
|
||||
help='Maximum number of charts in the image')
|
||||
parser.add_argument('--backend', default="default", type=str, choices=backends,
|
||||
help="Choose one of computation backends: "
|
||||
"default: automatically (by default), "
|
||||
"openvino: Intel's Deep Learning Inference Engine (https://software.intel.com/openvino-toolkit), "
|
||||
"opencv: OpenCV implementation, "
|
||||
"vkcom: VKCOM, "
|
||||
"cuda: CUDA, "
|
||||
"webnn: WebNN")
|
||||
parser.add_argument('--target', default="cpu", type=str, choices=targets,
|
||||
help="Choose one of target computation devices: "
|
||||
"cpu: CPU target (by default), "
|
||||
"opencl: OpenCL, "
|
||||
"opencl_fp16: OpenCL fp16 (half-float precision), "
|
||||
"vpu: VPU, "
|
||||
"vulkan: Vulkan, "
|
||||
"cuda: CUDA, "
|
||||
"cuda_fp16: CUDA fp16 (half-float preprocess)")
|
||||
|
||||
args, _ = parser.parse_known_args()
|
||||
add_preproc_args(args.zoo, parser, 'mcc', 'mcc')
|
||||
parser = argparse.ArgumentParser(parents=[parser],
|
||||
description='''
|
||||
To run:
|
||||
Default (compute new CCM):
|
||||
python color_correction_model.py --input=path/to/your/input/image --query=path/to/query/image
|
||||
DNN model:
|
||||
python color_correction_model.py mcc --input=path/to/your/input/image --query=path/to/query/image
|
||||
Using pre-computed CCM:
|
||||
python color_correction_model.py --ccm_file=path/to/ccm_output.yaml --query=path/to/query/image
|
||||
|
||||
Model path can also be specified using --model argument. And config path can be specified using --config.
|
||||
''', formatter_class=argparse.RawTextHelpFormatter)
|
||||
return parser.parse_args(func_args)
|
||||
|
||||
def process_frame(frame, detector, num_charts):
|
||||
if not detector.process(frame, num_charts):
|
||||
return None
|
||||
|
||||
checkers = detector.getListColorChecker()
|
||||
src = checkers[0].getChartsRGB(False)
|
||||
|
||||
return src
|
||||
|
||||
def main(func_args=None):
|
||||
args = get_args_parser(func_args)
|
||||
|
||||
if not (0 <= args.chart_type <= 2):
|
||||
raise ValueError("chartType must be 0, 1, or 2")
|
||||
|
||||
# Validate arguments based on whether using pre-computed CCM
|
||||
if args.ccm_file:
|
||||
if not args.query:
|
||||
print("[ERROR] Query image path must be provided when using pre-computed CCM.")
|
||||
return -1
|
||||
else:
|
||||
if not args.input:
|
||||
print("[ERROR] Input image path must be provided when computing new CCM.")
|
||||
return -1
|
||||
|
||||
# Read query image
|
||||
query_image = None
|
||||
if args.query:
|
||||
query_image = cv.imread(findFile(args.query))
|
||||
if query_image is None:
|
||||
print("[ERROR] Unable to read query image.")
|
||||
return -1
|
||||
|
||||
if os.getenv('OPENCV_SAMPLES_DATA_PATH') is not None:
|
||||
try:
|
||||
args.model = findModel(args.model, args.sha1)
|
||||
args.config = findModel(args.config, args.config_sha1)
|
||||
except:
|
||||
print("[WARN] Model file not provided, using default detector. Pass model using --model and config using --config to use dnn based detector.\n\n")
|
||||
args.model = None
|
||||
args.config = None
|
||||
else:
|
||||
args.model = None
|
||||
args.config = None
|
||||
print("[WARN] Model file not provided, using default detector. Pass model using --model and config using --config to use dnn based detector. Or, set OPENCV_SAMPLES_DATA_PATH environment variable.\n\n")
|
||||
|
||||
# Create color correction model
|
||||
model = cv.ccm.ColorCorrectionModel()
|
||||
|
||||
if args.ccm_file:
|
||||
# Load CCM from YAML file
|
||||
fs = cv.FileStorage(args.ccm_file, cv.FileStorage_READ)
|
||||
if not fs.isOpened():
|
||||
print(f"[ERROR] Unable to open CCM file: {args.ccm_file}")
|
||||
return -1
|
||||
model.read(fs.getNode("ColorCorrectionModel"))
|
||||
fs.release()
|
||||
print(f"Loaded CCM from file: {args.ccm_file}")
|
||||
else:
|
||||
# Read input image for computing new CCM
|
||||
image = cv.imread(findFile(args.input))
|
||||
if image is None:
|
||||
print("[ERROR] Unable to read input image.")
|
||||
return -1
|
||||
|
||||
# Create color checker detector
|
||||
if args.model and args.config:
|
||||
# Load the DNN from TensorFlow model
|
||||
engine = cv.dnn.ENGINE_AUTO
|
||||
if args.backend != "default" or args.target != "cpu":
|
||||
engine = cv.dnn.ENGINE_CLASSIC
|
||||
net = cv.dnn.readNetFromTensorflow(args.model, args.config, engine)
|
||||
net.setPreferableBackend(get_backend_id(args.backend))
|
||||
net.setPreferableTarget(get_target_id(args.target))
|
||||
|
||||
detector = cv.mcc_CCheckerDetector.create(net)
|
||||
print("Detecting checkers using neural network.")
|
||||
else:
|
||||
detector = cv.mcc_CCheckerDetector.create()
|
||||
print("Detecting checkers using default method (no DNN).")
|
||||
|
||||
detector.setColorChartType(args.chart_type)
|
||||
|
||||
# Process image to detect color checker
|
||||
src = process_frame(image, detector, args.num_charts)
|
||||
if src is None:
|
||||
print("No chart detected in the input image!")
|
||||
return -1
|
||||
|
||||
print("Actual colors:", src)
|
||||
|
||||
# Convert to double and normalize
|
||||
src = src.astype(np.float64) / 255.0
|
||||
|
||||
# Create and configure color correction model
|
||||
model = cv.ccm.ColorCorrectionModel(src, cv.ccm.COLORCHECKER_MACBETH)
|
||||
model.setCcmType(cv.ccm.CCM_LINEAR)
|
||||
model.setDistance(cv.ccm.DISTANCE_CIE2000)
|
||||
model.setLinearization(cv.ccm.LINEARIZATION_GAMMA)
|
||||
model.setLinearizationGamma(2.2)
|
||||
|
||||
# Compute color correction matrix
|
||||
ccm = model.compute()
|
||||
print("Computed CCM Matrix:\n", ccm)
|
||||
print("Loss:", model.getLoss())
|
||||
|
||||
# Save model parameters to YAML file
|
||||
fs = cv.FileStorage("ccm_output.yaml", cv.FileStorage_WRITE)
|
||||
model.write(fs)
|
||||
fs.release()
|
||||
print("Model parameters saved to ccm_output.yaml")
|
||||
|
||||
# Set query image for correction if not provided
|
||||
if query_image is None:
|
||||
print("[WARN] No query image provided, applying color correction on input image")
|
||||
query_image = image.copy()
|
||||
|
||||
# Apply correction to query image
|
||||
calibrated_image = np.empty_like(query_image)
|
||||
model.correctImage(query_image, calibrated_image)
|
||||
|
||||
cv.imshow("Original Image", query_image)
|
||||
cv.imshow("Corrected Image", calibrated_image)
|
||||
cv.waitKey(0)
|
||||
|
||||
return 0
|
||||
|
||||
if __name__ == "__main__":
|
||||
main()
|
||||
Reference in New Issue
Block a user