安裝SDK

這邊需要：

1. Face Api的sdk - 讓呼叫起來比較容易。
2. pillow - 這個是python裡面要修改圖片非常方便的一個套件。

!pip install cognitive_face
!pip install pillow

設定好呼叫服務的環境

主要設定：

1. 用到的api endpoint
2. 使用到的api key
3. 測試的圖片
4. 其他要偵測的attribute

faceURI = "https://eastasia.api.cognitive.microsoft.com/face/v1.0/"
faceKey = ""

img_url = 'https://raw.githubusercontent.com/Microsoft/Cognitive-Face-Windows/master/Data/detection1.jpg'

attributes = ('age,gender')

import cognitive_face as CF

# Set URI and Key
CF.BaseUrl.set(faceURI)
CF.Key.set(faceKey)

呈現圖片Detect得到的結果

接下來是呼叫Detect并且把Detect的值顯示出來：

result = CF.face.detect(img_url, False, False, attributes)
print (result)

把圖片呈現出來

%matplotlib inline
import requests
from io import BytesIO
from matplotlib.pyplot import imshow
from PIL import Image, ImageDraw, ImageFont

# Get the image
response = requests.get(img_url)
img = Image.open(BytesIO(response.content))

imshow(img)

把圖片裡面人臉的部分框起來

lineWidth = 8
color="red"

if result is not None:
    draw = ImageDraw.Draw(img) 
    for currFace in result:
        faceRectangle = currFace['faceRectangle']
        left = faceRectangle['left']
        top = faceRectangle['top']
        width = faceRectangle['width']
        height = faceRectangle['height']
        draw.line([(left,top),(left+width,top)],fill=color, width=lineWidth)
        draw.line([(left+width,top),(left+width,top+height)],fill=color , width=lineWidth)
        draw.line([(left+width,top+height),(left, top+height)],fill=color , width=lineWidth)
        draw.line([(left,top+height),(left, top)],fill=color , width=lineWidth)

# show the image
imshow(img)

顯示圖片人物識別出來的信息

!curl https://raw.githubusercontent.com/openscenegraph/OpenSceneGraph-Data/master/fonts/arial.ttf -o arial.ttf
    
font = ImageFont.truetype("arial.ttf", 45)

if result is not None:
    draw = ImageDraw.Draw(img) 
    for currFace in result:
        faceRectangle = currFace['faceRectangle']
        left = faceRectangle['left']
        top = faceRectangle['top']
        width = faceRectangle['width']
        height = faceRectangle['height']
        draw.text([left, height + 150], 'Age:' + str(currFace['faceAttributes']['age']), font=font, fill=(0,0,255))
        
# show the image
imshow(img)