Method postprocess

src/CVObjectDetection.cpp:121–246

Remove the bounding boxes with low confidence using non-maxima suppression

// Remove the bounding boxes with low confidence using non-maxima suppression
void CVObjectDetection::postprocess(const cv::Size &frameDims, const std::vector<cv::Mat>& outs, size_t frameId)
{
    std::vector<int> classIds;
    std::vector<float> confidences;
    std::vector<cv::Rect> boxes;
    std::vector<int> objectIds;

    for (size_t i = 0; i < outs.size(); ++i)
    {
        // Scan through all the bounding boxes output from the network and keep only the
        // ones with high confidence scores. Assign the box's class label as the class
        // with the highest score for the box.
        float* data = (float*)outs[i].data;
        for (int j = 0; j < outs[i].rows; ++j, data += outs[i].cols)
        {
            cv::Mat scores = outs[i].row(j).colRange(5, outs[i].cols);
            cv::Point classIdPoint;
            double confidence;
            // Get the value and location of the maximum score
            cv::minMaxLoc(scores, 0, &confidence, 0, &classIdPoint);
            if (confidence > confThreshold)
            {
                int centerX = (int)(data[0] * frameDims.width);
                int centerY = (int)(data[1] * frameDims.height);
                int width = (int)(data[2] * frameDims.width);
                int height = (int)(data[3] * frameDims.height);
                int left = centerX - width / 2;
                int top = centerY - height / 2;

                classIds.push_back(classIdPoint.x);
                confidences.push_back((float)confidence);
                boxes.push_back(cv::Rect(left, top, width, height));
            }
        }
    }

    // Perform non maximum suppression to eliminate redundant overlapping boxes with
    // lower confidences
    std::vector<int> indices;
    cv::dnn::NMSBoxes(boxes, confidences, confThreshold, nmsThreshold, indices);

    // Pass boxes to SORT algorithm
    std::vector<cv::Rect> sortBoxes;
    for(auto box : boxes)
        sortBoxes.push_back(box);
    sort.update(sortBoxes, frameId, sqrt(pow(frameDims.width,2) + pow(frameDims.height, 2)), confidences, classIds);

    // Clear data vectors
    boxes.clear(); confidences.clear(); classIds.clear(); objectIds.clear();
    // Get SORT predicted boxes
    for(auto TBox : sort.frameTrackingResult){
        if(TBox.frame == frameId){
            boxes.push_back(TBox.box);
            confidences.push_back(TBox.confidence);
            classIds.push_back(TBox.classId);
            objectIds.push_back(TBox.id);
        }
    }