Added MaxWeight mean picking + fixed the algorithm selection

src/plugins/opencv/nodes/superpixels/mean.cpp

@@ -9,25 +9,48 @@
 
 namespace {
 
-struct MeanAttrs {
-	virtual ~MeanAttrs() = default;
+enum Mode {
+	kMean,
+	kMedian,
+	kMaxWeight
+};
 
+struct MeanAttrs {
 	dependency_graph::InAttr<possumwood::opencv::Frame> a_in;
 	dependency_graph::InAttr<possumwood::opencv::Frame> a_superpixels;
 	dependency_graph::InAttr<possumwood::Enum> a_mode;
 	dependency_graph::OutAttr<possumwood::opencv::Frame> a_out;
 
-	virtual std::function<float(int, int)> weightFn(dependency_graph::Values& data) const;
+	std::function<float(int, int)> weightFn(dependency_graph::Values& data) const;
 	void init(possumwood::Metadata& meta);
 
+	virtual const std::vector<std::pair<std::string, int>>& modes() const {
+		static std::vector<std::pair<std::string, int>> s_mode {
+			{"mean", kMean},
+			{"median", kMedian},
+		};
+
+		return s_mode;
+	}
+
 	dependency_graph::State compute(dependency_graph::Values& data) const;
 };
 
 struct WeightedMeanAttrs : public MeanAttrs {
 	dependency_graph::InAttr<possumwood::opencv::Frame> a_weights;
 
-	virtual std::function<float(int, int)> weightFn(dependency_graph::Values& data) const override;
+	std::function<float(int, int)> weightFn(dependency_graph::Values& data) const;
 	void init(possumwood::Metadata& meta);
+
+	virtual const std::vector<std::pair<std::string, int>>& modes() const override {
+		static std::vector<std::pair<std::string, int>> s_mode {
+			{"mean", kMean},
+			{"median", kMedian},
+			{"max weight", kMaxWeight},
+		};
+
+		return s_mode;
+	}
 };
 
 MeanAttrs s_meanAttrs;
@@ -59,16 +82,6 @@ void setFloat(cv::Mat& in, int row, int col, int channel, float val) {
 	}
 }
 
-enum Mode {
-	kMean,
-	kMedian,
-};
-
-static std::vector<std::pair<std::string, int>> s_mode {
-	{"mean", kMean},
-	{"median", kMedian},
-};
-
 class Mean {
 	public:
 		Mean() : m_val(0.0f), m_norm(0.0f) {
@@ -109,7 +122,8 @@ class Median {
 			if(m_weightsSum == 0.0f)
 				return 0.0f;
 
-			std::sort(m_val.begin(), m_val.end());
+			std::sort(m_val.begin(), m_val.end(),
+				[](const std::pair<float, float>& p1, const std::pair<float, float>& p2) { return p1.first < p2.first; });
 
 			auto it = m_val.begin();
 			float weight = 0.0f;
@@ -130,6 +144,35 @@ class Median {
 		float m_weightsSum;
 };
 
+class MaxWeight {
+	public:
+		MaxWeight() {
+		}
+
+		void add(float val, float weight) {
+			if(weight < 0.0f)
+				throw std::runtime_error("Negative weight in weights input!");
+
+			m_vals[val] += weight;
+		}
+
+		float operator*() const {
+			if(!m_vals.empty()) {
+				auto max = m_vals.begin();
+
+				for(auto it = m_vals.begin(); it != m_vals.end(); ++it)
+					if(max->second < it->second)
+						max = it;
+
+				return max->first;
+			}
+			return 0.0f;
+		}
+
+	private:
+		std::map<float, float> m_vals;
+};
+
 template<typename MEAN>
 cv::Mat process(const cv::Mat& in, const cv::Mat& superpixels, std::function<float(int, int)> weights = [](int, int) {return 1.0f;}) {
 	// first of all, get the maximum index of the superpixels
@@ -139,7 +182,7 @@ cv::Mat process(const cv::Mat& in, const cv::Mat& superpixels, std::function<flo
 			maxIndex = std::max(maxIndex, superpixels.at<int32_t>(row, col));
 
 	// make the right sized accumulator and norm array
-	std::vector<std::vector<Mean>> vals(in.channels(), std::vector<Mean>(maxIndex+1, Mean()));
+	std::vector<std::vector<MEAN>> vals(in.channels(), std::vector<MEAN>(maxIndex+1, MEAN()));
 
 	// and accumulate the values
 	for(int row=0; row<in.rows; ++row)
@@ -199,6 +242,8 @@ dependency_graph::State MeanAttrs::compute(dependency_graph::Values& data) const
 		out = process<Mean>(in, superpixels, weights);
 	else if(data.get(a_mode).intValue() == kMedian)
 		out = process<Median>(in, superpixels, weights);
+	else if(data.get(a_mode).intValue() == kMaxWeight)
+		out = process<MaxWeight>(in, superpixels, weights);
 	else
 		throw std::runtime_error("Unknown mode " + data.get(a_mode).value());
 
@@ -210,7 +255,7 @@ dependency_graph::State MeanAttrs::compute(dependency_graph::Values& data) const
 void MeanAttrs::init(possumwood::Metadata& meta) {
 	meta.addAttribute(a_in, "in", possumwood::opencv::Frame(), possumwood::AttrFlags::kVertical);
 	meta.addAttribute(a_superpixels, "superpixels", possumwood::opencv::Frame(), possumwood::AttrFlags::kVertical);
-	meta.addAttribute(a_mode, "mode", possumwood::Enum(s_mode.begin(), s_mode.end()));
+	meta.addAttribute(a_mode, "mode", possumwood::Enum(modes().begin(), modes().end()));
 	meta.addAttribute(a_out, "out", possumwood::opencv::Frame(), possumwood::AttrFlags::kVertical);
 
 	meta.addInfluence(a_in, a_out);