10
3

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?

More than 3 years have passed since last update.

インタラクティブな画像(2D•3D)や動画(2D•3D)を作成する(その1)

Posted at

 関わっているプロジェクトで、インタラクティブな画像(2D・3D-360°)や動画(2D・3D-360°)を作成する機会があり、その際の手順を備忘録として少しずつまとめていこうと思います。

今回やりたいこと

 今回は、手始めとして「2D画像をインタラクティブにする手順」についてまとめます。
 具体的には、Flickr からダウンロードした、以下の画像の中の2頭の犬をクリッカプルにし、クリックした際に何らかのアクションを実行できるようにします。
 イメージマップを使って実現させますが、その際に Free Online Image Map Generator などのツールでは、画像が多数になった場合や、今後対応する3D(360°)画像・動画では対応出来ないため使用しません。その代わり、Detectron2 を使って、犬の位置を自動検出し、その位置情報を使ってイメージマップを作成します。

DSC_0079
photo by Putneypics

手順

環境

アクションを設定する対象の座標を抽出

 座標の抽出は、Google Colaboratory 上で、Detectron2 を使って行います。

 ※ Google Colaboratoryの基本的な使い方は、「Google Colabの知っておくべき使い方 – Google Colaboratoryのメリット・デメリットや基本操作のまとめ」 で分かりやすく説明されています。

###Detectron2のインストール
 Detectron2、その他必要なモジュールをインストールします。

google_colaboratory

!pip install -U torch torchvision cython
!pip install -U 'git+https://github.com/facebookresearch/fvcore.git' 'git+https://github.com/cocodataset/cocoapi.git#subdirectory=PythonAPI'
import torch, torchvision
torch.__version__
google_colaboratory
!git clone https://github.com/facebookresearch/detectron2 detectron2_repo
!pip install -e detectron2_repo

 イントール完了後、ランタイムを再起動する必要があります。再起動後に必要なライプラリーをimportします。

google_colaboratory

import detectron2
from detectron2.utils.logger import setup_logger
setup_logger()

import numpy as np
import cv2
import random
from google.colab.patches import cv2_imshow

from detectron2.engine import DefaultPredictor
from detectron2.config import get_cfg
from detectron2.data import MetadataCatalog

座標の取得

画像の読込み

 対象の画像を Flickr からダウンロードし、読込みます。

google_colaboratory

!wget https://live.staticflickr.com/3443/3702141833_7ea19dbd3d_z_d.jpg -O dog.jpg
img = cv2.imread("dog.jpg")

####画像内のオブジェクトを検出
 本来であれば、検出したいオブジェクトを学習させ推定を行うべきですが、今回は簡単のため学習済みモデルを使用しオブジェクトを検出します。

 ※ 独自データを学習させての推定は、動画対応の際に行いたいと思います。

google_colaboratory

cfg = get_cfg()
cfg.merge_from_file("./detectron2_repo/configs/COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x.yaml")
cfg.MODEL.ROI_HEADS.SCORE_THRESH_TEST = 0.7
cfg.MODEL.WEIGHTS = "detectron2://COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x/137849600/model_final_f10217.pkl"
predictor = DefaultPredictor(cfg)
outputs = predictor(img)

####検出したオブジェクトの中から、対象のマスクデータを抽出
 検出したデータの中から、今回対象にしている犬のマスクデータを取り出します。

google_colaboratory

target = MetadataCatalog.get(cfg.DATASETS.TRAIN[0]).thing_classes.index("dog")
classes = np.asarray(outputs["instances"].to("cpu").pred_classes)
masks = np.asarray(outputs["instances"].to("cpu").pred_masks)[classes==target].astype("uint8")

DOG1
img_mask_0.jpg
DOG2
img_mask_1.jpg

マスクデータから輪廓を抽出

google_colaboratory

contours = [ cv2.findContours(m, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)[0] for m in masks]

DOG1(右) + DOG2(左)
img_contours.jpg

輪郭の座標を取得

 輪郭の情報から、imagemapで使用するため、[x1,y1,x2,y2,...,xn,yn]の形で座標を取得します。

google_colaboratory

coords = [c.flatten() for contour in contours for c in contour]

DOG1

435,105,434,106,433,106,427,112,426,112,424,114,424,115,423,116,423,117,421,119,421,120,420,121,420,122,419,123,419,124,418,125,418,126,416,128,416,129,412,133,411,133,411,134,407,138,405,138,404,139,401,139,400,138,398,138,394,134,394,133,389,128,389,127,388,126,388,125,387,124,387,123,386,122,386,121,385,120,385,118,384,117,384,115,381,112,380,112,378,110,377,110,376,109,375,109,373,107,371,107,370,106,368,106,366,108,365,108,361,112,360,112,358,114,358,115,357,116,357,120,356,121,356,125,355,126,355,148,354,149,354,155,353,156,353,163,352,164,352,178,351,179,351,232,352,233,352,240,353,241,353,245,354,246,354,248,355,249,355,253,356,254,356,259,357,260,357,265,358,266,358,267,359,268,359,269,360,270,360,271,361,272,361,275,362,276,362,278,363,279,363,280,364,281,364,282,365,283,365,286,366,287,366,291,367,292,367,295,368,296,368,297,370,299,370,301,371,302,371,305,372,306,372,307,373,308,373,309,374,310,374,311,375,312,375,315,376,316,376,319,377,320,377,330,376,331,376,334,375,335,375,338,374,339,374,340,372,342,372,343,368,347,368,348,367,349,367,358,369,360,370,360,371,361,381,361,382,360,383,360,384,359,385,359,390,354,391,354,394,351,394,349,395,348,395,339,396,338,396,334,397,333,397,330,398,329,398,327,399,326,399,324,400,323,400,322,401,321,401,320,402,319,404,319,405,318,407,318,410,321,410,323,412,325,412,326,413,327,413,330,414,331,414,342,415,343,415,349,416,350,416,351,421,356,421,357,425,361,427,361,428,362,441,362,442,361,443,361,445,359,445,356,444,355,444,352,443,351,443,348,442,347,442,346,440,344,440,343,439,342,439,341,438,340,438,337,437,336,437,333,436,332,436,325,437,324,437,320,438,319,438,318,439,317,439,316,441,314,443,316,444,316,445,317,447,317,448,318,449,318,450,319,452,319,453,318,455,318,456,317,457,317,458,316,460,316,460,315,461,314,461,313,462,312,462,293,463,292,463,283,464,282,464,280,465,279,465,278,466,277,466,274,467,273,467,270,468,269,468,266,469,265,469,264,470,263,470,262,471,261,471,259,472,258,472,254,473,253,473,251,474,250,474,248,475,247,475,245,476,244,476,240,477,239,477,234,478,233,478,229,479,228,479,225,480,224,480,220,481,219,481,205,480,204,480,198,479,197,479,196,478,195,478,192,477,191,477,189,476,188,476,187,474,185,474,184,473,183,473,182,472,181,472,180,469,177,468,177,467,176,466,176,462,172,462,171,461,170,460,170,459,169,457,169,456,168,454,168,453,167,452,167,451,166,450,166,449,165,449,164,448,163,448,161,447,160,447,158,446,157,446,156,445,155,445,145,446,144,446,141,447,140,447,137,448,136,448,116,447,115,447,114,446,113,446,112,440,106,438,106,437,105
>*DOG2*

>```
184,19,183,20,180,20,179,21,177,21,176,22,174,22,173,23,171,23,169,25,168,25,167,26,160,26,159,25,158,25,157,24,156,24,155,23,154,23,153,22,152,22,151,21,150,21,149,20,140,20,139,21,137,21,131,27,130,27,128,29,127,29,126,30,124,30,123,31,122,31,121,32,120,32,118,34,118,35,117,36,117,37,116,38,116,40,115,41,115,42,114,43,114,44,113,45,113,47,112,48,112,51,113,52,113,53,114,54,114,55,121,62,121,63,122,64,122,65,123,66,123,69,124,70,124,74,125,75,125,77,124,78,124,89,123,90,123,94,122,95,122,99,121,100,121,103,120,104,120,107,119,108,119,111,118,112,118,116,117,117,117,121,116,122,116,125,115,126,115,128,114,129,114,132,113,133,113,137,112,138,112,156,113,157,113,167,114,168,114,171,115,172,115,176,116,177,116,190,117,191,117,218,118,219,118,221,119,222,119,224,120,225,120,226,121,227,121,228,122,229,122,230,123,231,123,233,124,234,124,236,125,237,125,238,127,240,127,241,128,242,128,246,129,247,129,254,130,255,130,266,129,267,129,274,128,275,128,279,127,280,127,283,126,284,126,289,125,290,125,299,124,300,124,310,123,311,123,323,122,324,122,329,121,330,121,334,120,335,120,341,119,342,119,348,120,349,120,352,121,353,121,354,122,355,122,356,123,357,123,359,126,362,127,362,128,363,129,363,130,364,132,364,133,365,137,365,138,366,143,366,144,367,147,367,148,368,152,368,153,369,155,369,156,368,157,368,158,367,159,367,160,366,162,366,163,365,165,365,166,364,167,364,169,362,169,361,170,360,170,359,171,358,171,357,172,356,172,355,173,354,173,353,174,352,174,351,175,350,175,349,176,348,176,346,177,345,177,341,178,340,178,337,179,336,179,333,180,332,180,330,181,329,181,328,182,327,182,326,183,325,183,324,184,323,186,325,186,326,187,327,187,328,188,329,188,331,189,332,189,336,190,337,190,347,191,348,191,353,192,354,192,359,193,360,193,362,194,363,194,364,196,366,197,366,198,367,200,367,201,368,215,368,216,367,217,367,218,366,219,366,220,365,221,365,225,361,226,362,229,362,230,363,232,363,233,364,235,364,236,365,239,365,240,364,243,364,244,363,245,363,246,362,247,362,249,360,249,359,255,353,256,353,257,352,258,352,261,349,261,348,262,347,262,344,263,343,263,339,264,338,264,334,265,333,265,324,266,323,266,318,265,317,265,307,264,306,264,299,263,298,263,285,262,284,262,269,261,268,261,264,260,263,260,260,259,259,259,257,258,256,258,255,257,254,257,253,256,252,256,251,252,247,251,247,248,244,248,243,247,242,247,241,246,240,246,239,244,237,244,236,242,234,242,233,241,232,241,231,240,230,240,223,239,222,239,215,238,214,238,204,237,203,237,201,236,200,236,198,235,197,235,194,234,193,234,190,233,189,233,188,232,187,232,185,231,184,231,183,230,182,230,181,229,180,229,176,228,175,228,172,227,171,227,166,226,165,226,163,225,162,225,161,224,160,224,158,223,157,223,154,222,153,222,147,221,146,221,140,220,139,220,136,219,135,219,134,218,133,218,131,217,130,217,126,216,125,216,100,217,99,217,84,216,83,216,73,215,72,215,68,214,67,214,64,213,63,214,62,214,60,215,59,215,58,216,57,216,56,217,55,217,53,218,52,218,51,219,50,219,45,220,44,220,35,219,34,219,33,218,32,217,32,207,22,206,22,205,21,204,21,203,20,201,20,200,19

画像のインタラクティブ化

 Imagemap、jQueryjQuery-rwdImageMaps で画像をインタラクティブかつレスポンシブにします。詳細については、他で分かりやすく説明されているサイトがありますので、そちらを参照いていただければと思います。

 ※ 実際のコード は以下のようになります。 コードサンプルは こちら

html

<img src="https://live.staticflickr.com/3443/3702141833_7ea19dbd3d_z.jpg" width="640" height="426" alt="Dog" usemap="#dogmap"/>
<map name="dogmap">
    <area class="dog1" shape="poly" coords="DOG1の座標" href="#" />
    <area class="dog2" shape="poly" coords="DOG2の座標" href="#" />
</map>
css
img[usemap] {
    max-width: 100%;
    height: auto;
}
script
$(document).ready(function(e) {
    $('img[usemap]').rwdImageMaps();
});
$(".dog1").on("click", function(e){
    e.preventDefault();
    alert('DOG1');
});
$(".dog2").on("click", function(e){
    e.preventDefault();
    alert('DOG2');
});

まとめ

 今回の方法で、インタラクティブな2D画像を、ほぼ自動で簡単に作成することが出来るようになります。
 今後は、この方法を発展させたインタラクティブな3D画像や動画を作成する手順を、出来るだけ簡潔にまとめていきたいと考えています。

10
3
0

Register as a new user and use Qiita more conveniently

  1. You get articles that match your needs
  2. You can efficiently read back useful information
  3. You can use dark theme
What you can do with signing up
10
3

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?