Merge remote-tracking branch 'upstream/3.4' into merge-3.4

cmake/OpenCVDetectCUDA.cmake

@@ -82,13 +82,14 @@ if(CUDA_FOUND)
 
   message(STATUS "CUDA detected: " ${CUDA_VERSION})
 
-  set(_generations "Fermi" "Kepler" "Maxwell" "Pascal" "Volta" "Turing")
+  set(_generations "Fermi" "Kepler" "Maxwell" "Pascal" "Volta" "Turing" "Ampere")
   set(_arch_fermi   "2.0")
   set(_arch_kepler  "3.0;3.5;3.7")
   set(_arch_maxwell "5.0;5.2")
   set(_arch_pascal  "6.0;6.1")
   set(_arch_volta   "7.0")
   set(_arch_turing  "7.5")
+  set(_arch_ampere  "8.0")
   if(NOT CMAKE_CROSSCOMPILING)
     list(APPEND _generations "Auto")
   endif()
@@ -163,6 +164,8 @@ if(CUDA_FOUND)
     set(__cuda_arch_bin ${_arch_volta})
   elseif(CUDA_GENERATION STREQUAL "Turing")
     set(__cuda_arch_bin ${_arch_turing})
+  elseif(CUDA_GENERATION STREQUAL "Ampere")
+    set(__cuda_arch_bin ${_arch_ampere})
   elseif(CUDA_GENERATION STREQUAL "Auto")
     ocv_detect_native_cuda_arch(_nvcc_res _nvcc_out)
     if(NOT _nvcc_res EQUAL 0)
@@ -180,7 +183,13 @@ if(CUDA_FOUND)
       ocv_detect_native_cuda_arch(_nvcc_res _nvcc_out)
       if(NOT _nvcc_res EQUAL 0)
         message(STATUS "Automatic detection of CUDA generation failed. Going to build for all known architectures.")
-        set(__cuda_arch_bin "5.3 6.2 7.2")
+        # TX1 (5.3) TX2 (6.2) Xavier (7.2) V100 (7.0)
+        ocv_filter_available_architecture(__cuda_arch_bin
+            5.3
+            6.2
+            7.2
+            7.0
+        )
       else()
         set(__cuda_arch_bin "${_nvcc_out}")
       endif()
@@ -193,6 +202,7 @@ if(CUDA_FOUND)
           ${_arch_pascal}
           ${_arch_volta}
           ${_arch_turing}
+          ${_arch_ampere}
       )
     endif()
   endif()

doc/opencv.bib

@@ -346,7 +346,8 @@
   year = {2003},
   pages = {363--370},
   publisher = {Springer},
-  url = {https://arxiv.org/pdf/1808.01752}
+  url = {https://doi.org/10.1007/3-540-45103-X_50},
+  doi = {10.1007/3-540-45103-X_50}
 }
 @inproceedings{Farsiu03,
   author = {Farsiu, Sina and Robinson, Dirk and Elad, Michael and Milanfar, Peyman},

modules/dnn/src/darknet/darknet_io.cpp

@@ -797,7 +797,7 @@ namespace cv {
                         int classes = getParam<int>(layer_params, "classes", -1);
                         int num_of_anchors = getParam<int>(layer_params, "num", -1);
                         float thresh = getParam<float>(layer_params, "thresh", 0.2);
-                        float nms_threshold = getParam<float>(layer_params, "nms_threshold", 0.4);
+                        float nms_threshold = getParam<float>(layer_params, "nms_threshold", 0.0);
                         float scale_x_y = getParam<float>(layer_params, "scale_x_y", 1.0);
 
                         std::string anchors_values = getParam<std::string>(layer_params, "anchors", std::string());

modules/dnn/src/layers/elementwise_layers.cpp

@@ -763,8 +763,11 @@ struct MishFunctor : public BaseFunctor
         {
             for( int i = 0; i < len; i++ )
             {
+                // Use fast approximation introduced in https://github.com/opencv/opencv/pull/17200
                 float x = srcptr[i];
-                dstptr[i] = x * tanh(log(1.0f + exp(x)));
+                float eX = exp(std::min(x, 20.f));
+                float n = (eX + 2) * eX;
+                dstptr[i] = (x * n) / (n + 2);
             }
         }
     }