Fix ONNX deconvolution

modules/dnn/src/onnx/onnx_importer.cpp

@@ -622,6 +622,14 @@ void ONNXImporter::populateNet(Net dstNet)
                     layerParams.set("adj_h", (outH - kernelH) % strideY);
                 }
             }
+            else if (layerParams.has("output_padding"))
+            {
+                const DictValue& adj_pad = layerParams.get("output_padding");
+                if (adj_pad.size() != 2)
+                    CV_Error(Error::StsNotImplemented, "Deconvolution3D layer is not supported");
+                layerParams.set("adj_w", adj_pad.get<int>(1));
+                layerParams.set("adj_h", adj_pad.get<int>(0));
+            }
         }
         else if (layer_type == "Transpose")
         {

modules/dnn/test/test_onnx_importer.cpp

@@ -100,6 +100,7 @@ TEST_P(Test_ONNX_layers, Deconvolution)
     testONNXModels("two_deconvolution");
     testONNXModels("deconvolution_group");
     testONNXModels("deconvolution_output_shape");
+    testONNXModels("deconv_adjpad_2d");
 }
 
 TEST_P(Test_ONNX_layers, Dropout)