3 * Copyright (C) 2013 Miguel Casas-Sanchez <miguelecasassanchez@gmail.com>
5 * Permission is hereby granted, free of charge, to any person obtaining a
6 * copy of this software and associated documentation files (the "Software"),
7 * to deal in the Software without restriction, including without limitation
8 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
9 * and/or sell copies of the Software, and to permit persons to whom the
10 * Software is furnished to do so, subject to the following conditions:
12 * The above copyright notice and this permission notice shall be included in
13 * all copies or substantial portions of the Software.
15 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
18 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
20 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
21 * DEALINGS IN THE SOFTWARE.
23 * Alternatively, the contents of this file may be used under the
24 * GNU Lesser General Public License Version 2.1 (the "LGPL"), in
25 * which case the following provisions apply instead of the ones
28 * This library is free software; you can redistribute it and/or
29 * modify it under the terms of the GNU Library General Public
30 * License as published by the Free Software Foundation; either
31 * version 2 of the License, or (at your option) any later version.
33 * This library is distributed in the hope that it will be useful,
34 * but WITHOUT ANY WARRANTY; without even the implied warranty of
35 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
36 * Library General Public License for more details.
38 * You should have received a copy of the GNU Library General Public
39 * License along with this library; if not, write to the
40 * Free Software Foundation, Inc., 51 Franklin St, Fifth Floor,
41 * Boston, MA 02110-1301, USA.
45 * SECTION:element-grabcut
48 * This element is a wrapper around OpenCV grabcut implementation. GrabCut is an
49 * image segmentation method based on graph cuts technique. It can be seen as a
50 * way of fine-grain segmenting the image from some FG and BG "seed" areas. The
51 * OpenCV implementation follows the article [1].
52 * The "seed" areas are taken in this element from either an input bounding box
53 * coming from a face detection, or from alpha channel values. The input box is
54 * taken from a "face" event such as the one generated from the 'facedetect'
55 * element. The Alpha channel values should be one of the following (cv.hpp):
57 * GC_BGD = 0, //!< background
58 * GC_FGD = 1, //!< foreground
59 * GC_PR_BGD = 2, //!< most probably background
60 * GC_PR_FGD = 3 //!< most probably foreground
62 * with values over GC_PR_FGD interpreted as GC_PR_FGD. IN CASE OF no alpha mask
63 * input (all 0's or all 1's), the 'GstOpenCvFaceDetect-face' downstream event
64 * is used to create a bbox of PR_FG elements. If both foreground alpha
65 * is not specified and there is no face detection, nothing is done.
67 * [1] C. Rother, V. Kolmogorov, and A. Blake, "GrabCut: Interactive foreground
68 * extraction using iterated graph cuts, ACM Trans. Graph., vol. 23, pp. 309–314,
72 * <title>Example launch line</title>
74 * gst-launch-1.0 --gst-debug=grabcut=4 v4l2src device=/dev/video0 ! videoconvert ! grabcut ! videoconvert ! video/x-raw,width=320,height=240 ! ximagesink
76 * Another example launch line
78 * gst-launch-1.0 --gst-debug=grabcut=4 v4l2src device=/dev/video0 ! videoconvert ! facedetect display=0 ! videoconvert ! grabcut test-mode=true ! videoconvert ! video/x-raw,width=320,height=240 ! ximagesink
87 #include "gstgrabcut.h"
90 #include <opencv2/imgproc/imgproc_c.h>
92 #include <opencv2/imgproc/imgproc.hpp>
93 GST_DEBUG_CATEGORY_STATIC (gst_grabcut_debug);
94 #define GST_CAT_DEFAULT gst_grabcut_debug
97 /* Filter signals and args */
111 #define DEFAULT_TEST_MODE FALSE
112 #define DEFAULT_SCALE 1.6
114 G_DEFINE_TYPE (GstGrabcut, gst_grabcut, GST_TYPE_OPENCV_VIDEO_FILTER);
115 static GstStaticPadTemplate sink_factory = GST_STATIC_PAD_TEMPLATE ("sink",
118 GST_STATIC_CAPS (GST_VIDEO_CAPS_MAKE ("RGBA")));
120 static GstStaticPadTemplate src_factory = GST_STATIC_PAD_TEMPLATE ("src",
123 GST_STATIC_CAPS (GST_VIDEO_CAPS_MAKE ("RGBA")));
126 static void gst_grabcut_set_property (GObject * object, guint prop_id,
127 const GValue * value, GParamSpec * pspec);
128 static void gst_grabcut_get_property (GObject * object, guint prop_id,
129 GValue * value, GParamSpec * pspec);
131 static GstFlowReturn gst_grabcut_transform_ip (GstOpencvVideoFilter * filter,
132 GstBuffer * buf, IplImage * img);
133 static gboolean gst_grabcut_set_caps (GstOpencvVideoFilter * filter,
134 gint in_width, gint in_height, gint in_depth, gint in_channels,
135 gint out_width, gint out_height, gint out_depth, gint out_channels);
137 static void gst_grabcut_release_all_pointers (GstGrabcut * filter);
139 static gboolean gst_grabcut_stop (GstBaseTransform * basesrc);
140 static void compose_matrix_from_image (CvMat * output, IplImage * input);
142 static int initialise_grabcut (struct grabcut_params *GC, IplImage * image_c,
144 static int run_grabcut_iteration (struct grabcut_params *GC,
145 IplImage * image_c, CvMat * mask_c, CvRect * bbox);
146 static int run_grabcut_iteration2 (struct grabcut_params *GC,
147 IplImage * image_c, CvMat * mask_c, CvRect * bbox);
148 static int finalise_grabcut (struct grabcut_params *GC);
150 /* initialize the grabcut's class */
152 gst_grabcut_class_init (GstGrabcutClass * klass)
154 GObjectClass *gobject_class = (GObjectClass *) klass;
155 GstElementClass *element_class = GST_ELEMENT_CLASS (klass);
156 GstOpencvVideoFilterClass *cvbasefilter_class =
157 (GstOpencvVideoFilterClass *) klass;
158 GstBaseTransformClass *btrans_class = (GstBaseTransformClass *) klass;
160 gobject_class->set_property = gst_grabcut_set_property;
161 gobject_class->get_property = gst_grabcut_get_property;
163 btrans_class->stop = gst_grabcut_stop;
164 btrans_class->passthrough_on_same_caps = TRUE;
166 cvbasefilter_class->cv_trans_ip_func = gst_grabcut_transform_ip;
167 cvbasefilter_class->cv_set_caps = gst_grabcut_set_caps;
169 g_object_class_install_property (gobject_class, PROP_TEST_MODE,
170 g_param_spec_boolean ("test-mode", "test-mode",
171 "If true, the output RGB is overwritten with the segmented foreground. Alpha channel same as normal case ",
172 DEFAULT_TEST_MODE, (GParamFlags)
173 (GParamFlags) (G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS)));
175 g_object_class_install_property (gobject_class, PROP_SCALE,
176 g_param_spec_float ("scale", "scale",
177 "Grow factor for the face bounding box, if present", 1.0,
179 (GParamFlags) (G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS)));
181 gst_element_class_set_static_metadata (element_class,
182 "Grabcut-based image FG/BG segmentation", "Filter/Effect/Video",
183 "Runs Grabcut algorithm on input alpha. Values: BG=0, FG=1, PR_BG=2, PR_FGD=3; \
184 NOTE: larger values of alpha (notably 255) are interpreted as PR_FGD too. \n\
185 IN CASE OF no alpha mask input (all 0's or all 1's), the 'face' \
186 downstream event is used to create a bbox of PR_FG elements.\n\
187 IF nothing is present, then nothing is done.", "Miguel Casas-Sanchez <miguelecasassanchez@gmail.com>");
189 gst_element_class_add_static_pad_template (element_class, &src_factory);
190 gst_element_class_add_static_pad_template (element_class, &sink_factory);
194 /* initialize the new element
195 * instantiate pads and add them to element
196 * set pad calback functions
197 * initialize instance structure
200 gst_grabcut_init (GstGrabcut * filter)
202 filter->test_mode = DEFAULT_TEST_MODE;
203 filter->scale = DEFAULT_SCALE;
204 gst_opencv_video_filter_set_in_place (GST_OPENCV_VIDEO_FILTER (filter),
210 gst_grabcut_set_property (GObject * object, guint prop_id,
211 const GValue * value, GParamSpec * pspec)
213 GstGrabcut *grabcut = GST_GRABCUT (object);
217 grabcut->test_mode = g_value_get_boolean (value);
220 grabcut->scale = g_value_get_float (value);
223 G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
229 gst_grabcut_get_property (GObject * object, guint prop_id,
230 GValue * value, GParamSpec * pspec)
232 GstGrabcut *filter = GST_GRABCUT (object);
236 g_value_set_boolean (value, filter->test_mode);
239 g_value_set_float (value, filter->scale);
242 G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
247 /* GstElement vmethod implementations */
248 /* this function handles the link with other elements */
250 gst_grabcut_set_caps (GstOpencvVideoFilter * filter, gint in_width,
251 gint in_height, gint in_depth, gint in_channels, gint out_width,
252 gint out_height, gint out_depth, gint out_channels)
254 GstGrabcut *grabcut = GST_GRABCUT (filter);
257 size = cvSize (in_width, in_height);
259 /* If cvRGB is already allocated, it means there's a cap modification,
260 * so release first all the images. */
261 if (!grabcut->cvRGBin)
262 gst_grabcut_release_all_pointers (grabcut);
264 grabcut->cvRGBin = cvCreateImage (size, IPL_DEPTH_8U, 3);
266 grabcut->cvA = cvCreateImage (size, IPL_DEPTH_8U, 1);
267 grabcut->cvB = cvCreateImage (size, IPL_DEPTH_8U, 1);
268 grabcut->cvC = cvCreateImage (size, IPL_DEPTH_8U, 1);
269 grabcut->cvD = cvCreateImage (size, IPL_DEPTH_8U, 1);
271 grabcut->grabcut_mask = cvCreateMat (size.height, size.width, CV_8UC1);
272 cvZero (grabcut->grabcut_mask);
273 initialise_grabcut (&(grabcut->GC), grabcut->cvRGBin, grabcut->grabcut_mask);
280 gst_grabcut_stop (GstBaseTransform * basesrc)
282 GstGrabcut *filter = GST_GRABCUT (basesrc);
284 if (filter->cvRGBin != NULL)
285 gst_grabcut_release_all_pointers (filter);
291 gst_grabcut_release_all_pointers (GstGrabcut * filter)
293 cvReleaseImage (&filter->cvRGBin);
295 cvReleaseImage (&filter->cvA);
296 cvReleaseImage (&filter->cvB);
297 cvReleaseImage (&filter->cvC);
298 cvReleaseImage (&filter->cvD);
300 finalise_grabcut (&(filter->GC));
304 gst_grabcut_transform_ip (GstOpencvVideoFilter * filter, GstBuffer * buffer,
307 GstGrabcut *gc = GST_GRABCUT (filter);
310 GstVideoRegionOfInterestMeta *meta;
311 meta = gst_buffer_get_video_region_of_interest_meta (buffer);
313 gc->facepos.x = (meta->x) - ((gc->scale - 1) * meta->w / 2);
314 gc->facepos.y = (meta->y) - ((gc->scale - 1) * meta->h / 2);
315 gc->facepos.width = meta->w * gc->scale * 0.9;
316 gc->facepos.height = meta->h * gc->scale * 1.1;
318 memset (static_cast<void*>(&(gc->facepos)), 0, sizeof (gc->facepos));
321 /* normally input should be RGBA */
322 cvSplit (img, gc->cvA, gc->cvB, gc->cvC, gc->cvD);
323 cvCvtColor (img, gc->cvRGBin, CV_BGRA2BGR);
324 compose_matrix_from_image (gc->grabcut_mask, gc->cvD);
326 /* Pass cvD to grabcut_mask for the graphcut stuff but that only if
327 really there is something in the mask! otherwise -->input bbox is
329 alphapixels = cvCountNonZero (gc->cvD);
330 if ((0 < alphapixels) && (alphapixels < (gc->width * gc->height))) {
331 GST_INFO ("running on mask");
332 run_grabcut_iteration (&(gc->GC), gc->cvRGBin, gc->grabcut_mask, NULL);
335 if ((abs (gc->facepos.width) > 2) && (abs (gc->facepos.height) > 2)) {
336 GST_INFO ("running on bbox (%d,%d),(%d,%d)", gc->facepos.x, gc->facepos.y,
337 gc->facepos.width, gc->facepos.height);
338 run_grabcut_iteration2 (&(gc->GC), gc->cvRGBin, gc->grabcut_mask,
341 GST_WARNING ("No face info present, skipping frame.");
346 /* if we want to display, just overwrite the output */
348 /* get only FG, PR_FG */
349 cvAndS (gc->grabcut_mask, cvRealScalar (1), gc->grabcut_mask, NULL);
350 /* (saturated) FG, PR_FG --> 255 */
351 cvConvertScale (gc->grabcut_mask, gc->grabcut_mask, 255.0, 0.0);
353 cvAnd (gc->grabcut_mask, gc->cvA, gc->cvA, NULL);
354 cvAnd (gc->grabcut_mask, gc->cvB, gc->cvB, NULL);
355 cvAnd (gc->grabcut_mask, gc->cvC, gc->cvC, NULL);
358 cvMerge (gc->cvA, gc->cvB, gc->cvC, gc->cvD, img);
362 cvPoint (gc->facepos.x, gc->facepos.y),
363 cvPoint (gc->facepos.x + gc->facepos.width,
364 gc->facepos.y + gc->facepos.height), CV_RGB (255, 0, 255), 1, 8, 0);
370 /* entry point to initialize the plug-in
371 * initialize the plug-in itself
372 * register the element factories and other features
375 gst_grabcut_plugin_init (GstPlugin * plugin)
377 /* debug category for fltering log messages
380 GST_DEBUG_CATEGORY_INIT (gst_grabcut_debug, "grabcut",
382 "Grabcut image segmentation on either input alpha or input bounding box");
384 return gst_element_register (plugin, "grabcut", GST_RANK_NONE,
389 compose_matrix_from_image (CvMat * output, IplImage * input)
393 for (x = 0; x < output->cols; x++) {
394 for (y = 0; y < output->rows; y++) {
395 CV_MAT_ELEM (*output, uchar, y, x) =
396 (cvGetReal2D (input, y, x) <= GC_PR_FGD) ? cvGetReal2D (input, y,
404 initialise_grabcut (struct grabcut_params *GC, IplImage * image_c,
407 GC->image = (void *) new Mat (cvarrToMat (image_c, false)); /* "true" refers to copydata */
408 GC->mask = (void *) new Mat (cvarrToMat (mask_c, false));
409 GC->bgdModel = (void *) new Mat (); /* "true" refers to copydata */
410 GC->fgdModel = (void *) new Mat ();
416 run_grabcut_iteration (struct grabcut_params *GC, IplImage * image_c,
417 CvMat * mask_c, CvRect * bbox)
419 ((Mat *) GC->image)->data = (uchar *) image_c->imageData;
420 ((Mat *) GC->mask)->data = mask_c->data.ptr;
422 if (cvCountNonZero (mask_c))
423 grabCut (*((Mat *) GC->image), *((Mat *) GC->mask), Rect (),
424 *((Mat *) GC->bgdModel), *((Mat *) GC->fgdModel), 1,
431 run_grabcut_iteration2 (struct grabcut_params *GC, IplImage * image_c,
432 CvMat * mask_c, CvRect * bbox)
434 ((Mat *) GC->image)->data = (uchar *) image_c->imageData;
435 ((Mat *) GC->mask)->data = mask_c->data.ptr;
436 grabCut (*((Mat *) GC->image), *((Mat *) GC->mask), *(bbox),
437 *((Mat *) GC->bgdModel), *((Mat *) GC->fgdModel), 1,
444 finalise_grabcut (struct grabcut_params *GC)
446 delete ((Mat *) GC->image);
447 delete ((Mat *) GC->mask);
448 delete ((Mat *) GC->bgdModel);
449 delete ((Mat *) GC->fgdModel);