--- /dev/null
+/**
+ * NNStreamer OpenCV Custom Filter Example: Scaler
+ *
+ * Copyright (C) 2018 Sangjung Woo <sangjung.woo@samsung.com>
+ *
+ * LICENSE: LGPL-2.1
+ *
+ * @file nnstreamer_customfilter_opencv_scaler.cc
+ * @date 22 Oct 2018
+ * @brief OpenCV Custom NNStreamer Filter Example: Scaler
+ * @author Sangjung Woo <sangjung.woo@samsung.com>
+ * @bug No known bugs
+ * @see nnstreamer_customfilter_example_scaler_allocator.c
+ *
+ * This example scales an input tensor of [N][input_h][input_w][M]
+ * to an ouput tensor of [N][output_h][output_w][M].
+ *
+ * The custom property is to be given as, "custom=[new-x]x[new-y]", where new-x and new-y are unsigned integers.
+ * E.g., custom=640x480
+ *
+ */
+
+#include "opencv2/opencv.hpp"
+#include "opencv2/highgui/highgui.hpp"
+
+#include <cassert>
+#include <glib.h>
+#include <tensor_filter_custom.h>
+#include <tensor_common.h>
+
+/**
+ * @brief Private data structure
+ */
+typedef struct _pt_data
+{
+ uint32_t in_height;
+ uint32_t in_width;
+ uint32_t out_height;
+ uint32_t out_width;
+} pt_data;
+
+/**
+ * @brief init callback of tensor_filter custom
+ */
+static void *
+pt_init (const GstTensorFilterProperties * prop)
+{
+ pt_data *data = g_new(pt_data, 1);
+ data->out_width = 0;
+ data->out_height = 0;
+
+ /* In case that custom property is given */
+ if (prop->custom_properties && strlen (prop->custom_properties) > 0) {
+ const char s[7] = "xX:_/ ";
+ gchar **strv = g_strsplit_set (prop->custom_properties, s, 3);
+ if (strv[0] != NULL) {
+ data->out_width = (uint32_t) g_ascii_strtoll (strv[0], NULL, 10);
+ } else {
+ data->out_width = 0;
+ }
+ if (strv[1] != NULL) {
+ data->out_height = (uint32_t) g_ascii_strtoll (strv[1], NULL, 10);
+ } else {
+ data->out_height = 0;
+ }
+ g_strfreev (strv);
+ }
+
+ return data;
+}
+
+/**
+ * @brief exit callback of tensor_filter custom
+ */
+static void
+pt_exit (void *private_data, const GstTensorFilterProperties * prop)
+{
+ pt_data *pdata = static_cast<pt_data *> (private_data);
+ assert (pdata);
+ g_free (pdata);
+}
+
+/**
+ * @brief setInputDimension callback of tensor_filter custom
+ */
+static int
+set_inputDim (void *private_data, const GstTensorFilterProperties * prop,
+ const GstTensorsInfo * in_info, GstTensorsInfo * out_info)
+{
+ pt_data *pdata = static_cast<pt_data *> (private_data);
+
+ assert (pdata);
+ assert (in_info);
+ assert (out_info);
+
+ out_info->num_tensors = 1;
+
+ for (int i = 0; i < NNS_TENSOR_RANK_LIMIT; i++)
+ out_info->info[0].dimension[i] = in_info->info[0].dimension[i];
+
+ /* Save width and height of an input tensor */
+ pdata->in_width = in_info->info[0].dimension[1];
+ pdata->in_height = in_info->info[0].dimension[2];
+
+ /* Update output dimension [1] and [2] with new-x, new-y */
+ if (pdata->out_width > 0)
+ out_info->info[0].dimension[1] = pdata->out_width;
+ if (pdata->out_height > 0)
+ out_info->info[0].dimension[2] = pdata->out_height;
+
+ out_info->info[0].type = in_info->info[0].type;
+ return 0;
+}
+
+/**
+ * @brief invoke-alloc callback of tensor_filter custom
+ */
+static int
+pt_allocate_invoke (void *private_data,
+ const GstTensorFilterProperties * prop, const GstTensorMemory * input,
+ GstTensorMemory * output)
+{
+ pt_data *pdata = static_cast<pt_data *> (private_data);
+ size_t in_size, out_size;
+ void* buffer;
+ cv::Mat img_src, img_dst;
+
+ assert (pdata);
+ assert (input);
+ assert (output);
+
+ in_size = gst_tensor_info_get_size (&prop->input_meta.info[0]);
+ out_size = gst_tensor_info_get_size (&prop->output_meta.info[0]);
+ buffer = g_malloc (in_size);
+ output[0].data = g_malloc (out_size);
+
+ /* Get Mat object from input tensor */
+ memcpy (buffer, input[0].data, in_size);
+ img_src = cv::Mat (pdata->in_height, pdata->in_width, CV_8UC3, buffer);
+ cv::cvtColor (img_src, img_src, CV_BGR2RGB);
+
+ /* Scale from the shape of input tensor to that of output tensor
+ * which is given as custom property */
+ cv::resize (img_src, img_dst, cv::Size(pdata->out_width, pdata->out_height),
+ 0, 0, CV_INTER_NN);
+
+ /* Convert Mat object to output tensor */
+ cv::cvtColor (img_dst, img_dst, CV_RGB2BGR);
+ memcpy (output[0].data, img_dst.data, out_size);
+
+ g_free(buffer);
+
+ return 0;
+}
+
+/**
+ * @brief tensor_filter custom subplugin definition
+ */
+static NNStreamer_custom_class NNStreamer_custom_body = {
+ .initfunc = pt_init,
+ .exitfunc = pt_exit,
+ .getInputDim = NULL,
+ .getOutputDim = NULL,
+ .setInputDim = set_inputDim,
+ .invoke = NULL,
+ .allocate_invoke = pt_allocate_invoke,
+};
+
+/* The dyn-loaded object */
+NNStreamer_custom_class *NNStreamer_custom = &NNStreamer_custom_body;
python checkScaledTensor.py testcase11.direct.log 640 480 testcase11.scaled.log 320 240 3
casereport 11 $? "Golden test comparison"
+# OpenCV Test
+# Test scaler using OpenCV (12, 13, 14)
+PATH_TO_MODEL="../../build/nnstreamer_example/custom_example_opencv/libnnstreamer_customfilter_opencv_scaler.so"
+gstTest "--gst-plugin-path=${PATH_TO_PLUGIN} videotestsrc num-buffers=1 ! video/x-raw,format=RGB,width=640,height=480,framerate=0/1 ! videoconvert ! video/x-raw, format=RGB ! tensor_converter ! tee name=t ! queue ! tensor_filter framework=\"custom\" model=\"${PATH_TO_MODEL}\" custom=\"640x480\" ! filesink location=\"testcase12.passthrough.log\" sync=true t. ! queue ! filesink location=\"testcase12.direct.log\" sync=true" 12
+compareAll testcase12.direct.log testcase12.passthrough.log 12
+
+gstTest "--gst-plugin-path=${PATH_TO_PLUGIN} videotestsrc num-buffers=1 ! video/x-raw,format=RGB,width=640,height=480,framerate=0/1 ! videoconvert ! video/x-raw, format=RGB ! tensor_converter ! tee name=t ! queue ! tensor_filter framework=\"custom\" model=\"${PATH_TO_MODEL}\" custom=\"320x240\" ! filesink location=\"testcase13.scaled.log\" sync=true t. ! queue ! filesink location=\"testcase13.direct.log\" sync=true" 13
+python checkScaledTensor.py testcase13.direct.log 640 480 testcase13.scaled.log 320 240 3
+casereport 13 $? "Golden test comparison"
+
+gstTest "--gst-plugin-path=${PATH_TO_PLUGIN} videotestsrc num-buffers=1 ! video/x-raw,format=RGB,width=640,height=480,framerate=0/1 ! videoconvert ! video/x-raw, format=RGB ! tensor_converter ! tee name=t ! queue ! tensor_filter framework=\"custom\" model=\"${PATH_TO_MODEL}\" custom=\"1920x1080\" ! filesink location=\"testcase14.scaled.log\" sync=true t. ! queue ! filesink location=\"testcase14.direct.log\" sync=true" 14
+python checkScaledTensor.py testcase14.direct.log 640 480 testcase14.scaled.log 1920 1080 3
+casereport 14 $? "Golden test comparison"
+
report
projects / platform / upstream / nnstreamer.git / commitdiff