1
0
mirror of https://github.com/opencv/opencv.git synced 2026-07-29 15:23:05 +04:00

Merge pull request #28510 from Anemptyship:optimize/resize-parallel-4.x

optimize(dnn): parallelize Resize layer implementation
This commit is contained in:
Alexander Smorkalov
2026-02-10 08:55:25 +03:00
committed by GitHub
2 changed files with 73 additions and 4 deletions
+63
View File
@@ -0,0 +1,63 @@
// This file is part of OpenCV project.
// It is subject to the license terms in the LICENSE file found in the top-level directory
// of this distribution and at http://opencv.org/license.html.
#include "perf_precomp.hpp"
namespace opencv_test {
struct Layer_Resize : public TestBaseWithParam<tuple<Backend, Target>>
{
void test_layer(const std::vector<int>& inpShape, int outH, int outW, const String& interp)
{
int backendId = get<0>(GetParam());
int targetId = get<1>(GetParam());
Mat input(inpShape, CV_32FC1);
randu(input, 0.f, 1.f);
Net net;
LayerParams lp;
lp.type = "Resize";
lp.name = "testLayer";
lp.set("interpolation", interp);
lp.set("width", outW);
lp.set("height", outH);
int id = net.addLayerToPrev(lp.name, lp.type, lp);
net.connect(0, 0, id, 0);
// warmup
{
net.setInputsNames({"data"});
net.setInput(input, "data");
net.setPreferableBackend(backendId);
net.setPreferableTarget(targetId);
Mat out = net.forward();
}
TEST_CYCLE()
{
Mat res = net.forward();
}
SANITY_CHECK_NOTHING();
}
};
PERF_TEST_P_(Layer_Resize, Resize_Upsample_Linear)
{
// N=4, C=64, H=64, W=64 -> 128x128 (x2 upsample)
// Common in segmentation/detection heads
test_layer({4, 64, 64, 64}, 128, 128, "opencv_linear");
}
PERF_TEST_P_(Layer_Resize, Resize_Downsample_Nearest)
{
// N=4, C=128, H=128, W=128 -> 64x64 (x0.5 downsample)
test_layer({4, 128, 128, 128}, 64, 64, "nearest");
}
INSTANTIATE_TEST_CASE_P(/**/, Layer_Resize, dnnBackendsAndTargets());
} // namespace opencv_test
+10 -4
View File
@@ -140,14 +140,20 @@ public:
{
// INTER_LINEAR Resize mode does not support INT8 inputs
InterpolationFlags mode = interpolation == "nearest" ? INTER_NEAREST : INTER_LINEAR;
for (size_t n = 0; n < inputs[0].size[0]; ++n)
{
for (size_t ch = 0; ch < inputs[0].size[1]; ++ch)
size_t nbatch = inputs[0].size[0];
size_t nch = inputs[0].size[1];
size_t total_planes = nbatch * nch;
parallel_for_(Range(0, (int)total_planes), [&](const Range& range){
for (int i = range.start; i < range.end; ++i)
{
int n = i / nch;
int ch = i % nch;
resize(getPlane(inp, n, ch), getPlane(out, n, ch),
Size(outWidth, outHeight), 0, 0, mode);
}
}
});
}
else if (interpolation == "nearest")
{