G-API-NG/Docs: Added a tutorial page on interactive face detection sample by dmatveev · Pull Request #16039 · opencv/opencv

dmatveev · 2019-12-02T20:42:55Z

Introduced a "--ser" option to run the pipeline serially for
benchmarking purposes
Reorganized sample code to better fit the documentation;
Fixed a couple of issues (mainly typos) in the public headers

force_builders=Custom,Custom Win,Custom Mac
build_image:Custom=ubuntu-openvino-2019r2.0:16.04
build_image:Custom Win=openvino-2019r2.0
build_image:Custom Mac=openvino-2019r2.0

test_modules:Custom=gapi
test_modules:Custom Win=gapi
test_modules:Custom Mac=gapi

buildworker:Custom=linux-1
test_opencl:Custom=ON
test_bigdata:Custom=1
test_filter:Custom=*

build_gapi_standalone:Linux x64=ade-0.1.1f
build_gapi_standalone:Win64=ade-0.1.1f
build_gapi_standalone:Mac=ade-0.1.1f
build_gapi_standalone:Linux x64 Debug=ade-0.1.1f
build_gapi_standalone:Custom Win=ade-0.1.1f

dmatveev · 2019-12-02T20:43:37Z

@dkurt @smirnov-alexey would you guys mind reviewing it? Thanks!

dmatveev · 2019-12-02T20:53:24Z

@OrestChura FYI

…mple - Introduced a "--ser" option to run the pipeline serially for benchmarking purposes - Reorganized sample code to better fit the documentation; - Fixed a couple of issues (mainly typos) in the public headers

dmatveev · 2019-12-04T10:23:47Z

@OrestChura @smirnov-alexey any updates with the review?

doc/tutorials/gapi/interactive_face_detection/interactive_face_detection.markdown

...les/cpp/tutorial_code/gapi/age_gender_emotion_recognition/age_gender_emotion_recognition.cpp

doc/tutorials/gapi/interactive_face_detection/interactive_face_detection.markdown

smirnov-alexey

Overall looks fine

dmatveev · 2019-12-05T06:53:47Z

@smirnov-alexey thanks for review, your comments are addressed. @dkurt is there anything to add? I am about to merge this already

dkurt · 2019-12-06T07:21:33Z

Do you mind if I will experiment a bit with DOT diagrams in doxygen? It seems to me we can reaplace a .png image to a text graph.

dmatveev · 2019-12-06T09:04:48Z

@dkurt then this needs to be done for all tutorials which are in progress :(

dkurt · 2019-12-06T10:22:45Z

So, the following code can generate similar graph but still remain some gaps (subgraph's border, label and crossing edge):

         * @dot
         *  digraph pipeline {
         *      node [shape=record fontname=Helvetica fontsize=10 style=filled color="#4c7aa4" fillcolor="#5b9bd5" fontcolor="white"];
         *      edge [color="#62a8e7"];
         *      splines=ortho;
         *      capture [label="Capture\nDecode" pos="1,2"];
         *      resize [label="Resize\nConvert"  pos="2,2"];
         *      detect [label="Detect faces"];
         *      show [label="Visualize\nDisplay"];
         *      temp_3 [style=invis shape=point width=0];
         *      subgraph cluster_1 {
         *        label="for each face";
         *        postproc_1 [label="Crop\nResize\nConvert"];
         *        postproc_2 [label="Crop\nResize\nConvert"];
         *        age_gender [label="Classify\nAge/gender"];
         *        emo [label="Classify\nEmotions"];
         *        temp_1 [style=invis shape=point width=0];
         *        temp_2 [style=invis shape=point width=0];
         *      }
         *      { rank=same; capture resize detect temp_1 postproc_1 age_gender temp_3 show }
         *      { rank=same; temp_2 postproc_2 emo }
         *      capture -> resize -> detect
         *      detect -> temp_1 [arrowhead=none]
         *      temp_1 -> postproc_1 -> age_gender
         *      temp_1 -> temp_2 [arrowhead=none]
         *      capture -> temp_2 [arrowhead=none]
         *      temp_2 -> postproc_2 -> emo
         *      capture -> postproc_1

         *      age_gender -> temp_3 [arrowhead=none]
         *      emo -> temp_3 [arrowhead=none]
         *      temp_3 -> show

         *      edge[style=invis];
         *      postproc_1 -> postproc_2
         *      age_gender -> emo
         *  }
         * @enddot

dmatveev · 2019-12-06T10:43:02Z

@dkurt https://graphviz.gitlab.io/_pages/Gallery/directed/cluster.html ? :)

dkurt · 2019-12-06T10:52:26Z

@dmatveev, thanks! Much more closer:

         * @dot
         *  digraph pipeline {
         *      node [shape=record fontname=Helvetica fontsize=10 style=filled color="#4c7aa4" fillcolor="#5b9bd5" fontcolor="white"];
         *      edge [color="#62a8e7"];
         *      splines=ortho;
         *      capture [label="Capture\nDecode"];
         *      resize [label="Resize\nConvert"];
         *      detect [label="Detect faces"];
         *      show [label="Visualize\nDisplay"];
         *      temp_3 [style=invis shape=point width=0];
         *      subgraph cluster_tmp {
         *        graph[style=dashed];
         *        postproc_1 [label="Crop\nResize\nConvert"];
         *        postproc_2 [label="Crop\nResize\nConvert"];
         *        age_gender [label="Classify\nAge/gender"];
         *        emo [label="Classify\nEmotions"];
         *        temp_1 [style=invis shape=point width=0];
         *        temp_2 [style=invis shape=point width=0];
         *        { rank=same; temp_2 postproc_2 emo }
         *        label="(for each face)";
         *      }
         *      { rank=same; capture resize detect temp_1 postproc_1 age_gender temp_3 show }
         *      capture -> resize -> detect
         *      detect -> temp_1 [arrowhead=none]
         *      temp_1 -> postproc_1 -> age_gender
         *      temp_1 -> temp_2 [arrowhead=none]
         *      capture -> temp_2 [arrowhead=none]
         *      temp_2 -> postproc_2 -> emo
         *      capture -> postproc_1

         *      age_gender -> temp_3 [arrowhead=none]
         *      emo -> temp_3 [arrowhead=none]
         *      temp_3 -> show

         *      edge[style=invis];
         *      postproc_1 -> postproc_2
         *      age_gender -> emo
         *  }
         * @enddot

dmatveev · 2019-12-06T11:14:06Z

@dkurt this is cool! can you please also put crop/resise->age/gender path to that cluster as well? Then it will be a 1:1

dkurt · 2019-12-07T11:07:45Z

@dmatveev, Done. May I ask you to try this graph instead of a png image?

source:

digraph pipeline {
  node [shape=record fontname=Helvetica fontsize=10 style=filled color="#4c7aa4" fillcolor="#5b9bd5" fontcolor="white"];
  edge [color="#62a8e7"];
  splines=ortho;

  rankdir = LR;
  subgraph cluster_0 {
    color=invis;
    capture [label="Capture\nDecode"];
    resize [label="Resize\nConvert"];
    detect [label="Detect faces"];
    capture -> resize -> detect
  }

  subgraph cluster_1 {
    graph[style=dashed];

    subgraph cluster_2 {
      color=invis;
      temp_4 [style=invis shape=point width=0];
      postproc_1 [label="Crop\nResize\nConvert"];
      age_gender [label="Classify\nAge/gender"];
      postproc_1 -> age_gender [constraint=true]
      temp_4 -> postproc_1 [constraint=none]
    }

    subgraph cluster_3 {
      color=invis;
      postproc_2 [label="Crop\nResize\nConvert"];
      emo [label="Classify\nEmotions"];
      postproc_2 -> emo [constraint=true]
    }
    label="(for each face)";
  }

  temp_1 [style=invis shape=point width=0];
  temp_2 [style=invis shape=point width=0];
  detect -> temp_1 [arrowhead=none]
  temp_1 -> postproc_1

  capture -> {temp_4, temp_2} [arrowhead=none constraint=false]
  temp_2 -> postproc_2

  temp_1 -> temp_2 [arrowhead=none constraint=false]

  temp_3 [style=invis shape=point width=0];
  show [label="Visualize\nDisplay"];

  {age_gender, emo} -> temp_3 [arrowhead=none]
  temp_3 -> show
}

dmatveev · 2019-12-09T12:03:35Z

@dkurt done, thanks! For some reason, however, in both Firefox and Chromium it looks like this for me now:

Clicking the arrow button at the right opens a valid svg diagram

dkurt · 2019-12-09T12:09:37Z

@dmatveev, Yeah, I also faced this problem. However it works fine in case of

cd opencv_build/doc/doxygen/html
python -m SimpleHTTPServer

and then go to localhost:8000

…xample

dmatveev · 2019-12-09T12:14:50Z

@dkurt I believe that local HTML documentation shouldn't require running a server to view that :) Worked for me with this change in Doxyfile.in.

INTERACTIVE_SVG        = NO

I couldn't find any other usages of DOT/SVG in the repo, so @alalek if this works...

dkurt · 2019-12-09T12:15:43Z

doc/Doxyfile.in

 DIRECTORY_GRAPH        = YES
 DOT_IMAGE_FORMAT       = svg
-INTERACTIVE_SVG        = YES
+INTERACTIVE_SVG        = NO


Great! I couldn't find it :)

dmatveev · 2019-12-09T12:23:41Z

@OrestChura I believe you will have to do something similar with #15942

doc/tutorials/gapi/table_of_content_gapi.markdown

dkurt

👍 Thanks!

dmatveev · 2019-12-10T08:18:19Z

Great, thanks!!

…ctive_face_detection * G-API-NG/Docs: Added a tutorial page on interactive face detection sample - Introduced a "--ser" option to run the pipeline serially for benchmarking purposes - Reorganized sample code to better fit the documentation; - Fixed a couple of issues (mainly typos) in the public headers * G-API-NG/Docs: Reflected meta-less compilation in new G-API tutorial * G-API-NG/Docs: Addressed review comments on Face Analytics Pipeline example

dmatveev added category: documentation Documentation fix or update category: g-api / gapi labels Dec 2, 2019

dmatveev added this to the 4.x milestone Dec 2, 2019

dmatveev requested review from dkurt and smirnov-alexey December 2, 2019 20:42

dmatveev added 2 commits December 3, 2019 22:15

G-API-NG/Docs: Reflected meta-less compilation in new G-API tutorial

6390b11

dmatveev force-pushed the dm/gapi_tutorial_interactive_face_detection branch from 229135f to 6390b11 Compare December 3, 2019 19:36