LoginSignup
5
4

More than 1 year has passed since last update.

Script TOPでMediaPipeやーる(TouchDesigner、Python3.7.2、OpenCV4.5.2、Windows10)

Last updated at Posted at 2021-06-07

はじめに

Script TOPを使ってMediaPipeをやりまーす

開発環境

  • TouchDesigner.2021.13610
  • Windows 10
  • Anaconda
  • OpenCV 4.5.2
  • Python 3.7.2

導入

1.Dialog->Textport and DATsからバージョンチェック
image.png

2.anacondaで同じ環境を作成

conda info -e
conda create -n td python=3.7.2
conda activate td
pip install opencv-python
pip install opencv-contrib-python
pip install mediapipe

3.Edit->Preferencesの設定
image.png

  • Python 64-bit Module Pathに作成したAnacondaの場所を指定(例:C:\Users\good_\anaconda3\envs\td\Lib\site-packages
  • Increment Filename on SaveをOff

4.TouchDesignerの再起動して、mediapipeがインポートできていればOK

5.Video Device Input(TOP)、Resolution(TOP)、Script(TOP)を配置します
image.png

6.Script(TOP)の中身

# me - this DAT
# scriptOp - the OP which is cooking
import cv2
import numpy as np
import mediapipe as mp

mp_hands = mp.solutions.hands
hands = mp_hands.Hands(max_num_hands=2,min_detection_confidence=0.3,min_tracking_confidence=0.3)

# press 'Setup Parameters' in the OP to call this function to re-create the parameters.
def onSetupParameters(scriptOp):
	return

# called whenever custom pulse parameter is pushed
def onPulse(par):
	return


def onCook(scriptOp):
	input_image = scriptOp.inputs[0].numpyArray()
	image = input_image * 255
	image = image.astype('uint8')
	image_height, image_width = image.shape[:2]

	image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
	results = hands.process(image)

	draw_image = image.copy()
	if results.multi_hand_landmarks is not None:
		for hand_landmarks, handedness in zip(results.multi_hand_landmarks,results.multi_handedness):
			for index, landmark in enumerate(hand_landmarks.landmark):
				if landmark.visibility < 0 or landmark.presence < 0:
					continue
				landmark_x = min(int(landmark.x * image_width), image_width - 1)
				landmark_y = min(int(landmark.y * image_height), image_height - 1)
				cv2.circle(draw_image, (landmark_x, landmark_y), 5, (0, 255, 0), 2)

	draw_image = cv2.cvtColor(draw_image, cv2.COLOR_RGB2BGR)
	scriptOp.copyNumpyArray(draw_image)
	return

カスタムパラメータで新しいページを作成しようとしましたが、なぜか反映されませんでした。。。

def onSetupParameters(scriptOp):
	page = scriptOp.appendCustomPage('Custom')
	p = page.appendInt('MaxNumHands', label='Max Num Hands')
	p[0].default = 2
	p = page.appendFloat('MinDetectionConfidence', label='Min Detection Confidence')
	p[0].default = 0.5
	p = page.appendFloat('MinTrackingConfidence', label='Min Tracking Confidence')
	p[0].default = 0.5
	return


def onCook(scriptOp):
	MaxNumHands = scriptOp.par.MaxNumHands.eval()
	MinDetectionConfidence = scriptOp.par.MinDetectionConfidence.eval()
	MinTrackingConfidence = scriptOp.par.MinTrackingConfidence.eval()

7.hand_landmarksのTable(DAT)を追加し、Script(TOP)を下記のように編集します。これで手のランドマークの値がTableに表示されます。

# me - this DAT
# scriptOp - the OP which is cooking
import cv2
import numpy as np
import mediapipe as mp

mp_hands = mp.solutions.hands
hands = mp_hands.Hands(max_num_hands=2,min_detection_confidence=0.3,min_tracking_confidence=0.3)

op_hand_landmarks = op('hand_landmarks')

# press 'Setup Parameters' in the OP to call this function to re-create the parameters.
def onSetupParameters(scriptOp):
	return

# called whenever custom pulse parameter is pushed
def onPulse(par):
	return


def onCook(scriptOp):
	input_image = scriptOp.inputs[0].numpyArray()
	image = input_image * 255
	image = image.astype('uint8')
	image_height, image_width = image.shape[:2]

	image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
	results = hands.process(image)

	draw_image = image.copy()
	op_hand_landmarks.clear()
	op_hand_landmarks.appendRow(['x', 'y', 'z'])
	if results.multi_hand_landmarks is not None:
		for hand_landmarks, handedness in zip(results.multi_hand_landmarks,results.multi_handedness):
			for index, landmark in enumerate(hand_landmarks.landmark):
				if landmark.visibility < 0 or landmark.presence < 0:
					continue
				landmark_x = min(int(landmark.x * image_width), image_width - 1)
				landmark_y = min(int(landmark.y * image_height), image_height - 1)
				cv2.circle(draw_image, (landmark_x, landmark_y), 5, (0, 255, 0), 2)
				op_hand_landmarks.appendRow([landmark.x, landmark.y, landmark.z])

	draw_image = cv2.cvtColor(draw_image, cv2.COLOR_RGB2BGR)
	scriptOp.copyNumpyArray(draw_image)
	return

8.Shpere(SOP)、Transform(SOP)、Geometry(COMP)を配置します。
image.png

9.GeometryのInstancingをONにし、Translate OPに作成したTable(DAT)を割り当てます。(Table(DAT)をGeometryへD&Dして、Translate OPを選択、Translate X, Y, ZにそれぞれTableのラベルx,y,zを割り当て)
image.png

10.Transform(SOP)でUniform Scaleの値を調整します。
image.png

お疲れ様でした。

Script TOPを使ったサンプルがあるので参考まで
image.png

image.png

5
4
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
5
4