FFmpeg
vf_find_rect.c
Go to the documentation of this file.
1 /*
2  * Copyright (c) 2014-2015 Michael Niedermayer <michaelni@gmx.at>
3  *
4  * This file is part of FFmpeg.
5  *
6  * FFmpeg is free software; you can redistribute it and/or modify
7  * it under the terms of the GNU General Public License as published by
8  * the Free Software Foundation; either version 2 of the License, or
9  * (at your option) any later version.
10  *
11  * FFmpeg is distributed in the hope that it will be useful,
12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14  * GNU General Public License for more details.
15  *
16  * You should have received a copy of the GNU General Public License along
17  * with FFmpeg; if not, write to the Free Software Foundation, Inc.,
18  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
19  */
20 
21 /**
22  * @todo switch to dualinput
23  */
24 
25 #include "libavutil/imgutils.h"
26 #include "libavutil/opt.h"
27 #include "internal.h"
28 
29 #include "lavfutils.h"
30 
31 #define MAX_MIPMAPS 5
32 
33 typedef struct FOCContext {
34  AVClass *class;
35  float threshold;
36  int mipmaps;
37  int xmin, ymin, xmax, ymax;
38  char *obj_filename;
39  int last_x, last_y;
43 } FOCContext;
44 
45 #define OFFSET(x) offsetof(FOCContext, x)
46 #define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM
47 static const AVOption find_rect_options[] = {
48  { "object", "object bitmap filename", OFFSET(obj_filename), AV_OPT_TYPE_STRING, {.str = NULL}, .flags = FLAGS },
49  { "threshold", "set threshold", OFFSET(threshold), AV_OPT_TYPE_FLOAT, {.dbl = 0.5}, 0, 1.0, FLAGS },
50  { "mipmaps", "set mipmaps", OFFSET(mipmaps), AV_OPT_TYPE_INT, {.i64 = 3}, 1, MAX_MIPMAPS, FLAGS },
51  { "xmin", "", OFFSET(xmin), AV_OPT_TYPE_INT, {.i64 = 0}, 0, INT_MAX, FLAGS },
52  { "ymin", "", OFFSET(ymin), AV_OPT_TYPE_INT, {.i64 = 0}, 0, INT_MAX, FLAGS },
53  { "xmax", "", OFFSET(xmax), AV_OPT_TYPE_INT, {.i64 = 0}, 0, INT_MAX, FLAGS },
54  { "ymax", "", OFFSET(ymax), AV_OPT_TYPE_INT, {.i64 = 0}, 0, INT_MAX, FLAGS },
55  { NULL }
56 };
57 
58 AVFILTER_DEFINE_CLASS(find_rect);
59 
61 {
62  static const enum AVPixelFormat pix_fmts[] = {
66  };
67 
68  return ff_set_common_formats(ctx, ff_make_format_list(pix_fmts));
69 }
70 
72 {
73  int x, y;
75  uint8_t *src, *dst;
76  if (!frame)
77  return NULL;
78 
79  frame->format = in->format;
80  frame->width = (in->width + 1) / 2;
81  frame->height = (in->height+ 1) / 2;
82 
83  if (av_frame_get_buffer(frame, 0) < 0) {
84  av_frame_free(&frame);
85  return NULL;
86  }
87  src = in ->data[0];
88  dst = frame->data[0];
89 
90  for(y = 0; y < frame->height; y++) {
91  for(x = 0; x < frame->width; x++) {
92  dst[x] = ( src[2*x+0]
93  + src[2*x+1]
94  + src[2*x+0 + in->linesize[0]]
95  + src[2*x+1 + in->linesize[0]]
96  + 2) >> 2;
97  }
98  src += 2*in->linesize[0];
99  dst += frame->linesize[0];
100  }
101  return frame;
102 }
103 
104 static float compare(const AVFrame *haystack, const AVFrame *obj, int offx, int offy)
105 {
106  int x,y;
107  int o_sum_v = 0;
108  int h_sum_v = 0;
109  int64_t oo_sum_v = 0;
110  int64_t hh_sum_v = 0;
111  int64_t oh_sum_v = 0;
112  float c;
113  int n = obj->height * obj->width;
114  const uint8_t *odat = obj ->data[0];
115  const uint8_t *hdat = haystack->data[0] + offx + offy * haystack->linesize[0];
116  int64_t o_sigma, h_sigma;
117 
118  for(y = 0; y < obj->height; y++) {
119  for(x = 0; x < obj->width; x++) {
120  int o_v = odat[x];
121  int h_v = hdat[x];
122  o_sum_v += o_v;
123  h_sum_v += h_v;
124  oo_sum_v += o_v * o_v;
125  hh_sum_v += h_v * h_v;
126  oh_sum_v += o_v * h_v;
127  }
128  odat += obj->linesize[0];
129  hdat += haystack->linesize[0];
130  }
131  o_sigma = n*oo_sum_v - o_sum_v*(int64_t)o_sum_v;
132  h_sigma = n*hh_sum_v - h_sum_v*(int64_t)h_sum_v;
133 
134  if (o_sigma == 0 || h_sigma == 0)
135  return 1.0;
136 
137  c = (n*oh_sum_v - o_sum_v*(int64_t)h_sum_v) / (sqrt(o_sigma)*sqrt(h_sigma));
138 
139  return 1 - fabs(c);
140 }
141 
143 {
144  AVFilterContext *ctx = inlink->dst;
145  FOCContext *foc = ctx->priv;
146 
147  if (foc->xmax <= 0)
148  foc->xmax = inlink->w - foc->obj_frame->width;
149  if (foc->ymax <= 0)
150  foc->ymax = inlink->h - foc->obj_frame->height;
151 
152  return 0;
153 }
154 
155 static float search(FOCContext *foc, int pass, int maxpass, int xmin, int xmax, int ymin, int ymax, int *best_x, int *best_y, float best_score)
156 {
157  int x, y;
158 
159  if (pass + 1 <= maxpass) {
160  int sub_x, sub_y;
161  search(foc, pass+1, maxpass, xmin>>1, (xmax+1)>>1, ymin>>1, (ymax+1)>>1, &sub_x, &sub_y, 2.0);
162  xmin = FFMAX(xmin, 2*sub_x - 4);
163  xmax = FFMIN(xmax, 2*sub_x + 4);
164  ymin = FFMAX(ymin, 2*sub_y - 4);
165  ymax = FFMIN(ymax, 2*sub_y + 4);
166  }
167 
168  for (y = ymin; y <= ymax; y++) {
169  for (x = xmin; x <= xmax; x++) {
170  float score = compare(foc->haystack_frame[pass], foc->needle_frame[pass], x, y);
171  if (score < best_score) {
172  best_score = score;
173  *best_x = x;
174  *best_y = y;
175  }
176  }
177  }
178  return best_score;
179 }
180 
182 {
183  AVFilterContext *ctx = inlink->dst;
184  FOCContext *foc = ctx->priv;
185  float best_score;
186  int best_x, best_y;
187  int i;
188 
189  foc->haystack_frame[0] = av_frame_clone(in);
190  for (i=1; i<foc->mipmaps; i++) {
191  foc->haystack_frame[i] = downscale(foc->haystack_frame[i-1]);
192  }
193 
194  best_score = search(foc, 0, 0,
195  FFMAX(foc->xmin, foc->last_x - 8),
196  FFMIN(foc->xmax, foc->last_x + 8),
197  FFMAX(foc->ymin, foc->last_y - 8),
198  FFMIN(foc->ymax, foc->last_y + 8),
199  &best_x, &best_y, 2.0);
200 
201  best_score = search(foc, 0, foc->mipmaps - 1, foc->xmin, foc->xmax, foc->ymin, foc->ymax,
202  &best_x, &best_y, best_score);
203 
204  for (i=0; i<MAX_MIPMAPS; i++) {
205  av_frame_free(&foc->haystack_frame[i]);
206  }
207 
208  if (best_score > foc->threshold) {
209  return ff_filter_frame(ctx->outputs[0], in);
210  }
211 
212  av_log(ctx, AV_LOG_DEBUG, "Found at %d %d score %f\n", best_x, best_y, best_score);
213  foc->last_x = best_x;
214  foc->last_y = best_y;
215 
217 
218  av_dict_set_int(&in->metadata, "lavfi.rect.w", foc->obj_frame->width, 0);
219  av_dict_set_int(&in->metadata, "lavfi.rect.h", foc->obj_frame->height, 0);
220  av_dict_set_int(&in->metadata, "lavfi.rect.x", best_x, 0);
221  av_dict_set_int(&in->metadata, "lavfi.rect.y", best_y, 0);
222 
223  return ff_filter_frame(ctx->outputs[0], in);
224 }
225 
227 {
228  FOCContext *foc = ctx->priv;
229  int i;
230 
231  for (i = 0; i < MAX_MIPMAPS; i++) {
232  av_frame_free(&foc->needle_frame[i]);
233  av_frame_free(&foc->haystack_frame[i]);
234  }
235 
236  if (foc->obj_frame)
237  av_freep(&foc->obj_frame->data[0]);
238  av_frame_free(&foc->obj_frame);
239 }
240 
242 {
243  FOCContext *foc = ctx->priv;
244  int ret, i;
245 
246  if (!foc->obj_filename) {
247  av_log(ctx, AV_LOG_ERROR, "object filename not set\n");
248  return AVERROR(EINVAL);
249  }
250 
251  foc->obj_frame = av_frame_alloc();
252  if (!foc->obj_frame)
253  return AVERROR(ENOMEM);
254 
255  if ((ret = ff_load_image(foc->obj_frame->data, foc->obj_frame->linesize,
256  &foc->obj_frame->width, &foc->obj_frame->height,
257  &foc->obj_frame->format, foc->obj_filename, ctx)) < 0)
258  return ret;
259 
260  if (foc->obj_frame->format != AV_PIX_FMT_GRAY8) {
261  av_log(ctx, AV_LOG_ERROR, "object image is not a grayscale image\n");
262  return AVERROR(EINVAL);
263  }
264 
265  foc->needle_frame[0] = av_frame_clone(foc->obj_frame);
266  for (i = 1; i < foc->mipmaps; i++) {
267  foc->needle_frame[i] = downscale(foc->needle_frame[i-1]);
268  if (!foc->needle_frame[i])
269  return AVERROR(ENOMEM);
270  }
271 
272  return 0;
273 }
274 
275 static const AVFilterPad foc_inputs[] = {
276  {
277  .name = "default",
278  .type = AVMEDIA_TYPE_VIDEO,
279  .config_props = config_input,
280  .filter_frame = filter_frame,
281  },
282  { NULL }
283 };
284 
285 static const AVFilterPad foc_outputs[] = {
286  {
287  .name = "default",
288  .type = AVMEDIA_TYPE_VIDEO,
289  },
290  { NULL }
291 };
292 
294  .name = "find_rect",
295  .description = NULL_IF_CONFIG_SMALL("Find a user specified object."),
296  .priv_size = sizeof(FOCContext),
297  .init = init,
298  .uninit = uninit,
300  .inputs = foc_inputs,
301  .outputs = foc_outputs,
302  .priv_class = &find_rect_class,
303 };
static float compare(const AVFrame *haystack, const AVFrame *obj, int offx, int offy)
Definition: vf_find_rect.c:104
#define NULL
Definition: coverity.c:32
This structure describes decoded (raw) audio or video data.
Definition: frame.h:308
AVOption.
Definition: opt.h:248
misc image utilities
AVFilter ff_vf_find_rect
Definition: vf_find_rect.c:293
#define OFFSET(x)
Definition: vf_find_rect.c:45
static int filter_frame(AVFilterLink *inlink, AVFrame *in)
Definition: vf_find_rect.c:181
AVFilterFormats * ff_make_format_list(const int *fmts)
Create a list of supported formats.
Definition: formats.c:294
const char * name
Pad name.
Definition: internal.h:60
int ff_filter_frame(AVFilterLink *link, AVFrame *frame)
Send a frame of data to the next filter.
Definition: avfilter.c:1091
uint8_t
#define av_cold
Definition: attributes.h:88
AVFrame * av_frame_alloc(void)
Allocate an AVFrame and set its fields to default values.
Definition: frame.c:190
static float search(FOCContext *foc, int pass, int maxpass, int xmin, int xmax, int ymin, int ymax, int *best_x, int *best_y, float best_score)
Definition: vf_find_rect.c:155
AVOptions.
#define FLAGS
Definition: vf_find_rect.c:46
Undefined Behavior In the C some operations are like signed integer dereferencing freed accessing outside allocated Undefined Behavior must not occur in a C it is not safe even if the output of undefined operations is unused The unsafety may seem nit picking but Optimizing compilers have in fact optimized code on the assumption that no undefined Behavior occurs Optimizing code based on wrong assumptions can and has in some cases lead to effects beyond the output of computations The signed integer overflow problem in speed critical code Code which is highly optimized and works with signed integers sometimes has the problem that often the output of the computation does not c
Definition: undefined.txt:32
AVFrame * obj_frame
Definition: vf_find_rect.c:40
static av_cold void uninit(AVFilterContext *ctx)
Definition: vf_find_rect.c:226
AVDictionary * metadata
metadata.
Definition: frame.h:594
static const AVFilterPad foc_outputs[]
Definition: vf_find_rect.c:285
static AVFrame * downscale(AVFrame *in)
Definition: vf_find_rect.c:71
#define MAX_MIPMAPS
Definition: vf_find_rect.c:31
#define av_log(a,...)
A filter pad used for either input or output.
Definition: internal.h:54
static const AVFilterPad foc_inputs[]
Definition: vf_find_rect.c:275
static const AVOption find_rect_options[]
Definition: vf_find_rect.c:47
#define src
Definition: vp8dsp.c:254
int width
Definition: frame.h:366
#define AV_LOG_ERROR
Something went wrong and cannot losslessly be recovered.
Definition: log.h:194
int ff_set_common_formats(AVFilterContext *ctx, AVFilterFormats *formats)
A helper for query_formats() which sets all links to the same list of formats.
Definition: formats.c:607
void av_frame_free(AVFrame **frame)
Free the frame and any dynamically allocated objects in it, e.g.
Definition: frame.c:203
#define NULL_IF_CONFIG_SMALL(x)
Return NULL if CONFIG_SMALL is true, otherwise the argument without modification. ...
Definition: internal.h:153
void * priv
private data for use by the filter
Definition: avfilter.h:353
#define AV_LOG_DEBUG
Stuff which is only useful for libav* developers.
Definition: log.h:215
#define FFMAX(a, b)
Definition: common.h:94
#define pass
Definition: fft_template.c:619
#define FFMIN(a, b)
Definition: common.h:96
planar YUV 4:2:0, 12bpp, full scale (JPEG), deprecated in favor of AV_PIX_FMT_YUV420P and setting col...
Definition: pixfmt.h:78
AVFormatContext * ctx
Definition: movenc.c:48
these buffered frames must be flushed immediately if a new input produces new the filter must not call request_frame to get more It must just process the frame or queue it The task of requesting more frames is left to the filter s request_frame method or the application If a filter has several the filter must be ready for frames arriving randomly on any input any filter with several inputs will most likely require some kind of queuing mechanism It is perfectly acceptable to have a limited queue and to drop frames when the inputs are too unbalanced request_frame For filters that do not use the this method is called when a frame is wanted on an output For a it should directly call filter_frame on the corresponding output For a if there are queued frames already one of these frames should be pushed If the filter should request a frame on one of its repeatedly until at least one frame has been pushed Return or at least make progress towards producing a frame
int ff_load_image(uint8_t *data[4], int linesize[4], int *w, int *h, enum AVPixelFormat *pix_fmt, const char *filename, void *log_ctx)
Load image from filename and put the resulting image in data.
Definition: lavfutils.c:24
AVFrame * av_frame_clone(const AVFrame *src)
Create a new frame that references the same data as src.
Definition: frame.c:553
static const AVFilterPad outputs[]
Definition: af_acontrast.c:203
int format
format of the frame, -1 if unknown or unset Values correspond to enum AVPixelFormat for video frames...
Definition: frame.h:381
int linesize[AV_NUM_DATA_POINTERS]
For video, size in bytes of each picture line.
Definition: frame.h:339
AVFrame * haystack_frame[MAX_MIPMAPS]
Definition: vf_find_rect.c:42
these buffered frames must be flushed immediately if a new input produces new the filter must not call request_frame to get more It must just process the frame or queue it The task of requesting more frames is left to the filter s request_frame method or the application If a filter has several inputs
static av_cold int init(AVFilterContext *ctx)
Definition: vf_find_rect.c:241
uint8_t pi<< 24) CONV_FUNC_GROUP(AV_SAMPLE_FMT_FLT, float, AV_SAMPLE_FMT_U8, uint8_t,(*(const uint8_t *) pi-0x80)*(1.0f/(1<< 7))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_DBL, double, AV_SAMPLE_FMT_U8, uint8_t,(*(const uint8_t *) pi-0x80)*(1.0/(1<< 7))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_S16, int16_t,(*(const int16_t *) pi >> 8)+0x80) CONV_FUNC_GROUP(AV_SAMPLE_FMT_FLT, float, AV_SAMPLE_FMT_S16, int16_t,*(const int16_t *) pi *(1.0f/(1<< 15))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_DBL, double, AV_SAMPLE_FMT_S16, int16_t,*(const int16_t *) pi *(1.0/(1<< 15))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_S32, int32_t,(*(const int32_t *) pi >> 24)+0x80) CONV_FUNC_GROUP(AV_SAMPLE_FMT_FLT, float, AV_SAMPLE_FMT_S32, int32_t,*(const int32_t *) pi *(1.0f/(1U<< 31))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_DBL, double, AV_SAMPLE_FMT_S32, int32_t,*(const int32_t *) pi *(1.0/(1U<< 31))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_FLT, float, av_clip_uint8(lrintf(*(const float *) pi *(1<< 7))+0x80)) CONV_FUNC_GROUP(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_FLT, float, av_clip_int16(lrintf(*(const float *) pi *(1<< 15)))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_FLT, float, av_clipl_int32(llrintf(*(const float *) pi *(1U<< 31)))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_DBL, double, av_clip_uint8(lrint(*(const double *) pi *(1<< 7))+0x80)) CONV_FUNC_GROUP(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_DBL, double, av_clip_int16(lrint(*(const double *) pi *(1<< 15)))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_DBL, double, av_clipl_int32(llrint(*(const double *) pi *(1U<< 31))))#define SET_CONV_FUNC_GROUP(ofmt, ifmt) static void set_generic_function(AudioConvert *ac){}void ff_audio_convert_free(AudioConvert **ac){if(!*ac) return;ff_dither_free(&(*ac) ->dc);av_freep(ac);}AudioConvert *ff_audio_convert_alloc(AVAudioResampleContext *avr, enum AVSampleFormat out_fmt, enum AVSampleFormat in_fmt, int channels, int sample_rate, int apply_map){AudioConvert *ac;int in_planar, out_planar;ac=av_mallocz(sizeof(*ac));if(!ac) return NULL;ac->avr=avr;ac->out_fmt=out_fmt;ac->in_fmt=in_fmt;ac->channels=channels;ac->apply_map=apply_map;if(avr->dither_method!=AV_RESAMPLE_DITHER_NONE &&av_get_packed_sample_fmt(out_fmt)==AV_SAMPLE_FMT_S16 &&av_get_bytes_per_sample(in_fmt) > 2){ac->dc=ff_dither_alloc(avr, out_fmt, in_fmt, channels, sample_rate, apply_map);if(!ac->dc){av_free(ac);return NULL;}return ac;}in_planar=ff_sample_fmt_is_planar(in_fmt, channels);out_planar=ff_sample_fmt_is_planar(out_fmt, channels);if(in_planar==out_planar){ac->func_type=CONV_FUNC_TYPE_FLAT;ac->planes=in_planar?ac->channels:1;}else if(in_planar) ac->func_type=CONV_FUNC_TYPE_INTERLEAVE;else ac->func_type=CONV_FUNC_TYPE_DEINTERLEAVE;set_generic_function(ac);if(ARCH_AARCH64) ff_audio_convert_init_aarch64(ac);if(ARCH_ARM) ff_audio_convert_init_arm(ac);if(ARCH_X86) ff_audio_convert_init_x86(ac);return ac;}int ff_audio_convert(AudioConvert *ac, AudioData *out, AudioData *in){int use_generic=1;int len=in->nb_samples;int p;if(ac->dc){av_log(ac->avr, AV_LOG_TRACE,"%d samples - audio_convert: %s to %s (dithered)\n", len, av_get_sample_fmt_name(ac->in_fmt), av_get_sample_fmt_name(ac->out_fmt));return ff_convert_dither(ac-> in
Describe the class of an AVClass context structure.
Definition: log.h:67
AVFILTER_DEFINE_CLASS(find_rect)
Filter definition.
Definition: avfilter.h:144
const char * name
Filter name.
Definition: avfilter.h:148
AVFilterLink ** outputs
array of pointers to output links
Definition: avfilter.h:350
static enum AVPixelFormat pix_fmts[]
Definition: libkvazaar.c:270
int av_frame_get_buffer(AVFrame *frame, int align)
Allocate new buffer(s) for audio or video data.
Definition: frame.c:337
int av_frame_make_writable(AVFrame *frame)
Ensure that the frame data is writable, avoiding data copy if possible.
Definition: frame.c:624
uint8_t * data[AV_NUM_DATA_POINTERS]
pointer to the picture/channel planes.
Definition: frame.h:322
The exact code depends on how similar the blocks are and how related they are to the and needs to apply these operations to the correct inlink or outlink if there are several Macros are available to factor that when no extra processing is inlink
planar YUV 4:2:0, 12bpp, (1 Cr & Cb sample per 2x2 Y samples)
Definition: pixfmt.h:66
Y , 8bpp.
Definition: pixfmt.h:74
float threshold
Definition: vf_find_rect.c:35
int av_dict_set_int(AVDictionary **pm, const char *key, int64_t value, int flags)
Convenience wrapper for av_dict_set that converts the value to a string and stores it...
Definition: dict.c:147
static int config_input(AVFilterLink *inlink)
Definition: vf_find_rect.c:142
static int query_formats(AVFilterContext *ctx)
Definition: vf_find_rect.c:60
An instance of a filter.
Definition: avfilter.h:338
int height
Definition: frame.h:366
#define av_freep(p)
char * obj_filename
Definition: vf_find_rect.c:38
internal API functions
Filter the word “frame” indicates either a video frame or a group of audio as stored in an AVFrame structure Format for each input and each output the list of supported formats For video that means pixel format For audio that means channel sample they are references to shared objects When the negotiation mechanism computes the intersection of the formats supported at each end of a all references to both lists are replaced with a reference to the intersection And when a single format is eventually chosen for a link amongst the remaining all references to the list are updated That means that if a filter requires that its input and output have the same format amongst a supported all it has to do is use a reference to the same list of formats query_formats can leave some formats unset and return AVERROR(EAGAIN) to cause the negotiation mechanism toagain later.That can be used by filters with complex requirements to use the format negotiated on one link to set the formats supported on another.Frame references ownership and permissions
AVPixelFormat
Pixel format.
Definition: pixfmt.h:64
int i
Definition: input.c:407
Miscellaneous utilities which make use of the libavformat library.
AVFrame * needle_frame[MAX_MIPMAPS]
Definition: vf_find_rect.c:41