add mobilenetv3-ssd (#1335)

* update

* Update mobilenetv3ssdlite.cpp

* Update mobilenetv3ssdlite.cpp

* Update mobilenetv3ssdlite.cpp

examples/mobilenetv3ssdlite.cpp

@@ -0,0 +1,178 @@
+// Tencent is pleased to support the open source community by making ncnn available.
+//
+// Copyright (C) 2018 THL A29 Limited, a Tencent company. All rights reserved.
+//
+// Licensed under the BSD 3-Clause License (the "License"); you may not use this file except
+// in compliance with the License. You may obtain a copy of the License at
+//
+// https://opensource.org/licenses/BSD-3-Clause
+//
+// Unless required by applicable law or agreed to in writing, software distributed
+// under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR
+// CONDITIONS OF ANY KIND, either express or implied. See the License for the
+// specific language governing permissions and limitations under the License.
+
+#include <stdio.h>
+#include <vector>
+#include <opencv2/core/core.hpp>
+#include <opencv2/highgui/highgui.hpp>
+#include <opencv2/imgproc/imgproc.hpp>
+
+#include "platform.h"
+#include "net.h"
+#if NCNN_VULKAN
+#include "gpu.h"
+#endif // NCNN_VULKAN
+
+template<class T>
+const T& clamp(const T&v, const T& lo, const T& hi)
+{
+    assert(!(hi < lo));
+    return v < lo ? lo : hi < v ? hi : v;
+}
+
+struct Object
+{
+    cv::Rect_<float> rect;
+    int label;
+    float prob;
+};
+
+static int detect_mobilenetv3(const cv::Mat& bgr, std::vector<Object>& objects)
+{
+    ncnn::Net mobilenetv3;
+
+#if NCNN_VULKAN
+    mobilenetv3.opt.use_vulkan_compute = true;
+#endif // NCNN_VULKAN
+
+    // converted ncnn model from https://github.com/ujsyehao/mobilenetv3-ssd
+    mobilenetv3.load_param("./mobilenetv3_ssdlite_voc.param");
+    mobilenetv3.load_model("./mobilenetv3_ssdlite_voc.bin");
+
+    const int target_size = 300;
+
+    int img_w = bgr.cols;
+    int img_h = bgr.rows;
+
+    ncnn::Mat in = ncnn::Mat::from_pixels_resize(bgr.data, ncnn::Mat::PIXEL_BGR2RGB, bgr.cols, bgr.rows, target_size, target_size);
+
+    const float mean_vals[3] = {123.675f, 116.28f, 103.53f};
+    const float norm_vals[3] = {1.0f, 1.0f, 1.0f};
+    in.substract_mean_normalize(mean_vals, norm_vals);
+
+    ncnn::Extractor ex = mobilenetv3.create_extractor();
+    ex.set_light_mode(true);
+    ex.set_num_threads(4);
+
+    ex.input("input", in);
+
+    ncnn::Mat out;
+    ex.extract("detection_out",out);
+
+//     printf("%d %d %d\n", out.w, out.h, out.c);
+    objects.clear();
+    for (int i=0; i<out.h; i++)
+    {
+        const float* values = out.row(i);
+
+        Object object;
+        object.label = values[0];
+        object.prob = values[1];
+
+        // filter out cross-boundary
+        float x1 = clamp(values[2] * target_size, 0.f, float(target_size - 1)) / target_size * img_w;
+        float y1 = clamp(values[3] * target_size, 0.f, float(target_size - 1)) / target_size * img_h;
+        float x2 = clamp(values[4] * target_size, 0.f, float(target_size - 1)) / target_size * img_w;
+        float y2 = clamp(values[5] * target_size, 0.f, float(target_size - 1)) / target_size * img_h;		
+
+        object.rect.x = x1;
+        object.rect.y = y1;
+        object.rect.width = x2 - x1;
+        object.rect.height = y2 - y1;
+
+        objects.push_back(object);
+    }
+
+    return 0;
+}
+
+static void draw_objects(const cv::Mat& bgr, const std::vector<Object>& objects)
+{
+    static const char* class_names[] = {"background",
+        "aeroplane", "bicycle", "bird", "boat",
+        "bottle", "bus", "car", "cat", "chair",
+        "cow", "diningtable", "dog", "horse",
+        "motorbike", "person", "pottedplant",
+        "sheep", "sofa", "train", "tvmonitor"};
+
+    cv::Mat image = bgr.clone();
+
+    for (size_t i = 0; i < objects.size(); i++)
+    {
+        if (objects[i].prob > 0.6) 
+	  {
+	    const Object& obj = objects[i];
+
+	    fprintf(stderr, "%d = %.5f at %.2f %.2f %.2f x %.2f\n", obj.label, obj.prob,
+		    obj.rect.x, obj.rect.y, obj.rect.width, obj.rect.height);
+
+	    cv::rectangle(image, obj.rect, cv::Scalar(255, 0, 0));
+
+	    char text[256];
+	    sprintf(text, "%s %.1f%%", class_names[obj.label], obj.prob * 100);
+
+	    int baseLine = 0;
+	    cv::Size label_size = cv::getTextSize(text, cv::FONT_HERSHEY_SIMPLEX, 0.5, 1, &baseLine);
+
+	    int x = obj.rect.x;
+	    int y = obj.rect.y - label_size.height - baseLine;
+            if (y < 0)
+	        y = 0;
+	    if (x + label_size.width > image.cols)
+		x = image.cols - label_size.width;
+
+	    cv::rectangle(image, cv::Rect(cv::Point(x, y), cv::Size(label_size.width, label_size.height + baseLine)),
+		          cv::Scalar(255, 255, 255), -1);
+
+	    cv::putText(image, text, cv::Point(x, y + label_size.height),
+		        cv::FONT_HERSHEY_SIMPLEX, 0.5, cv::Scalar(0, 0, 0));
+	  }
+    }
+
+    cv::imshow("image", image);
+    cv::waitKey(0);
+}
+
+int main(int argc, char** argv)
+{
+    if (argc != 2)
+    {
+        fprintf(stderr, "Usage: %s [imagepath]\n", argv[0]);
+        return -1;
+    }
+
+    const char* imagepath = argv[1];
+
+    cv::Mat m = cv::imread(imagepath, 1);
+    if (m.empty())
+    {
+        fprintf(stderr, "cv::imread %s failed\n", imagepath);
+        return -1;
+    }
+
+#if NCNN_VULKAN
+    ncnn::create_gpu_instance();
+#endif // NCNN_VULKAN
+
+    std::vector<Object> objects;
+    detect_mobilenetv3(m, objects);
+
+#if NCNN_VULKAN
+    ncnn::destroy_gpu_instance();
+#endif // NCNN_VULKAN
+
+    draw_objects(m, objects);
+
+    return 0;
+}

src/layer/priorbox.cpp

@@ -42,6 +42,8 @@ int PriorBox::load_param(const ParamDict& pd)
     step_width = pd.get(11, -233.f);
     step_height = pd.get(12, -233.f);
     offset = pd.get(13, 0.f);
+    step_mmdetection = pd.get(14, 0);
+    center_mmdetection = pd.get(15, 0);
 
     return 0;
 }
@@ -137,8 +139,12 @@ int PriorBox::forward(const std::vector<Mat>& bottom_blobs, std::vector<Mat>& to
     float step_h = step_height;
     if (step_w == -233)
         step_w = (float)image_w / w;
+        if (step_mmdetection)
+            step_w = ceil((float)image_w / w);
     if (step_h == -233)
         step_h = (float)image_h / h;
+        if (step_mmdetection)
+            step_h = ceil((float)image_h / h);
 
     int num_min_size = min_sizes.w;
     int num_max_size = max_sizes.w;
@@ -160,6 +166,11 @@ int PriorBox::forward(const std::vector<Mat>& bottom_blobs, std::vector<Mat>& to
 
         float center_x = offset * step_w;
         float center_y = offset * step_h + i * step_h;
+        if (center_mmdetection) 
+        {
+            center_x = offset * (step_w - 1);
+            center_y = offset * (step_h - 1) + i * step_h;
+        }
 
         for (int j = 0; j < w; j++)
         {

src/layer/priorbox.h

@@ -40,6 +40,8 @@ public:
     float step_width;
     float step_height;
     float offset;
+    bool step_mmdetection;
+    bool center_mmdetection;
 };
 
 } // namespace ncnn