Timelapse Sample.

base/CMakeLists.txt

@@ -517,7 +517,6 @@ target_include_directories ( aprapipes PRIVATE
 	${NVCODEC_INCLUDE_DIR}
 )
 
-
 # aprapipes Unit Tests
 
 IF (ENABLE_ARM64)
@@ -565,6 +564,35 @@ IF (ENABLE_CUDA)
 	ENDIF(NOT ENABLE_ARM64)
 ENDIF(ENABLE_CUDA)
 
+SET(SAMPLE_UT_FILES
+	test/utmain.cpp
+	test/test_utils.cpp
+	test/test_utils.h
+	.././samples/timelapse-sample/timelapse_summary_test.cpp
+	.././samples/create_thumbnail_from_mp4_video/test_generate_thumbnail_from_mp4_video.cpp
+	.././samples/face_detection_cpu/test_face_detection_cpu.cpp
+	.././samples/play_mp4_from_beginning/test_play_mp4_video_from_beginning.cpp
+	.././samples/relay-sample/relay_sample_test.cpp
+)
+
+SET(SAMPLE_CORE_FILES
+	.././samples/timelapse-sample/timelapse_summary.h
+	.././samples/timelapse-sample/timelapse_summary.cpp
+	.././samples/create_thumbnail_from_mp4_video/GenerateThumbnailsPipeline.cpp
+	.././samples/create_thumbnail_from_mp4_video/GenerateThumbnailsPipeline.h
+	.././samples/face_detection_cpu/face_detection_cpu.h
+	.././samples/face_detection_cpu/face_detection_cpu.cpp
+	.././samples/play_mp4_from_beginning/PlayMp4VideoFromBeginning.h
+	.././samples/play_mp4_from_beginning/PlayMp4VideoFromBeginning.cpp
+	.././samples/relay-sample/relay_sample.h
+	.././samples/relay-sample/relay_sample.cpp
+)
+
+SET(SAMPLE_SOURCE
+	${SAMPLE_CORE_FILES}
+	${SAMPLE_UT_FILES}
+)
+
 SET(UT_FILES
 	test/utmain.cpp
 	test/unit_tests.cpp
@@ -638,6 +666,8 @@ ENDIF(ENABLE_LINUX)
 
 add_executable(aprapipesut ${UT_FILES})
 
+add_executable(aprapipessampleut ${SAMPLE_SOURCE})
+
 IF(ENABLE_ARM64)
 	target_include_directories ( aprapipesut PRIVATE ${JETSON_MULTIMEDIA_LIB_INCLUDE} ${FFMPEG_ROOT} ${JPEG_INCLUDE_DIR})
 ENDIF(ENABLE_ARM64)
@@ -661,6 +691,41 @@ IF(ENABLE_LINUX)
 	)	
 ENDIF(ENABLE_LINUX)
 
+target_include_directories ( aprapipessampleut PRIVATE 
+${JETSON_MULTIMEDIA_LIB_INCLUDE}
+${FFMPEG_INCLUDE_DIRS} 
+${OpenCV_INCLUDE_DIRS} 
+${Boost_INCLUDE_DIRS}
+${LIBMP4_INC_DIR}
+${BARESIP_INC_DIR}
+${LIBRE_INC_DIR}
+${NVCODEC_INCLUDE_DIR}
+test
+)
+
+target_link_libraries(aprapipessampleut   
+  aprapipes  
+  ${JPEG_LIBRARIES}
+  ${LIBMP4_LIB}
+  ${OPENH264_LIB}
+  ${Boost_LIBRARIES}
+  ${FFMPEG_LIBRARIES}
+  ${OpenCV_LIBRARIES}
+  ${JETSON_LIBS}
+  ${NVCUDAToolkit_LIBS}
+  ${NVCODEC_LIB}
+  ${NVJPEGLIB_L4T}
+  ${CURSES_LIBRARIES}
+  ZXing::Core
+  ZXing::ZXing	
+  BZip2::BZip2
+  ZLIB::ZLIB
+  liblzma::liblzma
+  bigint::bigint
+  sfml-audio
+  whisper::whisper
+  )
+
 target_link_libraries(aprapipesut   
   aprapipes  
   ${GLEW_LIBRARIES}
@@ -727,3 +792,5 @@ install(
 
 install(DIRECTORY ${CMAKE_CURRENT_LIST_DIR}/include
         DESTINATION ${CMAKE_INSTALL_INCLUDEDIR}/aprapipes)
+
+add_subdirectory(${CMAKE_CURRENT_SOURCE_DIR}/../samples ${CMAKE_CURRENT_BINARY_DIR}/samples)

base/include/FaceDetectorXform.h

@@ -6,11 +6,14 @@
 class FaceDetectorXformProps : public ModuleProps
 {
 public:
-    FaceDetectorXformProps(double _scaleFactor = 1.0, float _confidenceThreshold = 0.5) : scaleFactor(_scaleFactor), confidenceThreshold(_confidenceThreshold)
+    FaceDetectorXformProps(double _scaleFactor = 1.0, float _confidenceThreshold = 0.5, std::string _Face_Detection_Configuration= "./data/assets/deploy.prototxt", std::string _Face_Detection_Weights= "./data/assets/res10_300x300_ssd_iter_140000_fp16.caffemodel")
+        : scaleFactor(_scaleFactor), confidenceThreshold(_confidenceThreshold), FACE_DETECTION_CONFIGURATION(_Face_Detection_Configuration), FACE_DETECTION_WEIGHTS(_Face_Detection_Weights)
     {
     }
     double scaleFactor;
     float confidenceThreshold;
+    std::string FACE_DETECTION_CONFIGURATION;
+	std::string FACE_DETECTION_WEIGHTS;
 
     size_t getSerializeSize()
     {

base/include/FacialLandmarksCV.h

@@ -81,4 +81,5 @@ class FacialLandmarkCV : public Module
 	FacialLandmarkCVProps mProp;
 	bool handlePropsChange(frame_sp& frame);
 	std::string mOutputPinId1;
+	std::string rawFramePinId;
 };

base/include/MotionVectorExtractor.h

@@ -6,18 +6,15 @@ class MvExtractDetailAbs;
 class DetailFfmpeg;
 class DetailOpenH264;
 
-class MotionVectorExtractorProps : public ModuleProps
-{
+class MotionVectorExtractorProps : public ModuleProps {
 public:
-	enum MVExtractMethod
-	{
-		FFMPEG,
-		OPENH264
-	};
+	enum MVExtractMethod { FFMPEG, OPENH264 };
 
-	MotionVectorExtractorProps(MVExtractMethod _MVExtractMethod = MVExtractMethod::FFMPEG, bool _sendDecodedFrame = false, int _motionVectorThreshold = 2) : MVExtract(_MVExtractMethod), sendDecodedFrame(_sendDecodedFrame), motionVectorThreshold(_motionVectorThreshold)
-	{
-	}
+	MotionVectorExtractorProps(
+		MVExtractMethod _MVExtractMethod = MVExtractMethod::FFMPEG,
+		bool _sendDecodedFrame = false, int _motionVectorThreshold = 2)
+		: MVExtract(_MVExtractMethod), sendDecodedFrame(_sendDecodedFrame),
+		motionVectorThreshold(_motionVectorThreshold) {}
 
 	size_t getSerializeSize()
 	{
@@ -26,26 +23,26 @@ class MotionVectorExtractorProps : public ModuleProps
 	bool sendDecodedFrame = false;
 	int motionVectorThreshold;
 	MVExtractMethod MVExtract = MVExtractMethod::FFMPEG;
+
 private:
 	friend class boost::serialization::access;
 
 	template <class Archive>
-	void serialize(Archive& ar, const unsigned int version)
-	{
+	void serialize(Archive& ar, const unsigned int version) {
 		ar& boost::serialization::base_object<ModuleProps>(*this);
 		ar& sendDecodedFrame;
 		ar& motionVectorThreshold;
 	}
 };
 
-class MotionVectorExtractor : public Module
-{
+class MotionVectorExtractor : public Module {
 public:
 	MotionVectorExtractor(MotionVectorExtractorProps _props);
 	virtual ~MotionVectorExtractor() {};
 	bool init();
 	bool term();
 	void setProps(MotionVectorExtractorProps& props);
+
 protected:
 	bool process(frame_container& frame);
 	bool validateInputPins();
@@ -54,8 +51,9 @@ class MotionVectorExtractor : public Module
 	bool processSOS(frame_sp& frame);
 	void setMetadata(frame_sp metadata);
 	bool handlePropsChange(frame_sp& frame);
+
 private:
 	boost::shared_ptr<MvExtractDetailAbs> mDetail;
 	framemetadata_sp rawOutputMetadata;
 	bool mShouldTriggerSOS = true;
-};
+};

base/src/FaceDetectorXform.cpp

@@ -11,6 +11,7 @@
 #include "AIPExceptions.h"
 #include "ApraFaceInfo.h"
 #include "FaceDetectsInfo.h"
+#include "Overlay.h"
 
 class FaceDetectorXform::Detail
 {
@@ -34,13 +35,12 @@ class FaceDetectorXform::Detail
 	framemetadata_sp mOutputMetadata;
 	FaceDetectorXformProps mProps;
 	std::string mOutputPinId;
+	std::string rawFramePinId;
 	cv::Mat mInputImg;
 	int mFrameType;
 	cv::dnn::Net network;
 	cv::Mat inputBlob;
 	cv::Mat detection;
-	const std::string FACE_DETECTION_CONFIGURATION = "./data/assets/deploy.prototxt";
-	const std::string FACE_DETECTION_WEIGHTS = "./data/assets/res10_300x300_ssd_iter_140000_fp16.caffemodel";
 	// scalar with mean values which are subtracted from channels.
 	// Values are intended to be in (mean-R, mean-G, mean-B) order if image has BGR ordering and swapRB is true.
 	const cv::Scalar meanValuesRGB = cv::Scalar({104., 177.0, 123.0});
@@ -75,7 +75,7 @@ bool FaceDetectorXform::validateInputPins()
 
 bool FaceDetectorXform::validateOutputPins()
 {
-	if (getNumberOfOutputPins() != 1)
+	if (getNumberOfOutputPins() > 2)
 	{
 		LOG_ERROR << "<" << getId() << ">::validateOutputPins size is expected to be 1. Actual<" << getNumberOfOutputPins() << ">";
 		return false;
@@ -88,11 +88,12 @@ void FaceDetectorXform::addInputPin(framemetadata_sp &metadata, string &pinId)
 	Module::addInputPin(metadata, pinId);
 	mDetail->mOutputMetadata = framemetadata_sp(new FrameMetadata(FrameMetadata::FACEDETECTS_INFO));
 	mDetail->mOutputPinId = addOutputPin(mDetail->mOutputMetadata);
+	mDetail->rawFramePinId = addOutputPin(metadata);
 }
 
 bool FaceDetectorXform::init()
 {
-	mDetail->network = cv::dnn::readNetFromCaffe(mDetail->FACE_DETECTION_CONFIGURATION, mDetail->FACE_DETECTION_WEIGHTS);
+	mDetail->network = cv::dnn::readNetFromCaffe(mDetail->mProps.FACE_DETECTION_CONFIGURATION, mDetail->mProps.FACE_DETECTION_WEIGHTS);
 	if (mDetail->network.empty())
 	{
 		LOG_ERROR << "Failed to load network with the given settings. Please check the loaded parameters.";
@@ -121,11 +122,14 @@ bool FaceDetectorXform::process(frame_container &frames)
 	mDetail->inputBlob = cv::dnn::blobFromImage(mDetail->mInputImg, mDetail->mProps.scaleFactor, cv::Size(mDetail->mInputImg.cols, mDetail->mInputImg.rows),
 												mDetail->meanValuesRGB, false, false);
 	mDetail->network.setInput(mDetail->inputBlob, "data");
-
+	
 	mDetail->detection = mDetail->network.forward("detection_out");
 
 	cv::Mat detectionMatrix(mDetail->detection.size[2], mDetail->detection.size[3], CV_32F, mDetail->detection.ptr<float>());
 
+	std::vector<RectangleOverlay> rectangleOverlays;
+    CompositeOverlay compositeOverlay;
+
 	for (int i = 0; i < detectionMatrix.rows; i++)
 	{
 		float confidence = detectionMatrix.at<float>(i, 2);
@@ -136,20 +140,42 @@ bool FaceDetectorXform::process(frame_container &frames)
 		}
 
 		mDetail->faceInfo.x1 = detectionMatrix.at<float>(i, 3) * mDetail->mInputImg.cols;
-		mDetail->faceInfo.y2 = detectionMatrix.at<float>(i, 4) * mDetail->mInputImg.rows;
+		mDetail->faceInfo.y1 = detectionMatrix.at<float>(i, 4) * mDetail->mInputImg.rows;
 		mDetail->faceInfo.x2 = detectionMatrix.at<float>(i, 5) * mDetail->mInputImg.cols;
-		mDetail->faceInfo.y1 = detectionMatrix.at<float>(i, 6) * mDetail->mInputImg.rows;
+		mDetail->faceInfo.y2 = detectionMatrix.at<float>(i, 6) * mDetail->mInputImg.rows;
 		mDetail->faceInfo.score = confidence;
 
 		mDetail->faces.emplace_back(mDetail->faceInfo);
+
+		RectangleOverlay rectangleOverlay;
+		rectangleOverlay.x1 = mDetail->faceInfo.x1;
+		rectangleOverlay.y1 = mDetail->faceInfo.y1;
+		rectangleOverlay.x2 = mDetail->faceInfo.x2;
+		rectangleOverlay.y2 = mDetail->faceInfo.y2;
+		rectangleOverlays.push_back(rectangleOverlay);
 	}
 
+	for (auto &rectangleOverlay : rectangleOverlays) {
+          compositeOverlay.add(&rectangleOverlay);
+	}
+
+	auto rawFrame = frames.cbegin()->second;
+	frames.insert(make_pair(mDetail->rawFramePinId, rawFrame));
+
 	mDetail->faceDetectsInfo.faces = mDetail->faces;
-	auto outFrame = makeFrame(mDetail->faceDetectsInfo.getSerializeSize());
-	mDetail->faceDetectsInfo.serialize(outFrame->data(), mDetail->faceDetectsInfo.getSerializeSize());
-	frames.insert(make_pair(mDetail->mOutputPinId, outFrame));
+
+	if (rectangleOverlays.size() > 0) {
+		DrawingOverlay drawingOverlay;
+		drawingOverlay.add(&compositeOverlay);
+		auto mvSize = drawingOverlay.mGetSerializeSize();  
+		auto outFrame = makeFrame(mvSize, mDetail->mOutputPinId);
+		drawingOverlay.serialize(outFrame);
+        frames.insert(make_pair(mDetail->mOutputPinId, outFrame));
+    }
+
 	mDetail->faces.clear();
 	mDetail->faceDetectsInfo.faces.clear();
+
 	send(frames);
 	return true;
 }