CODEWITHSUNDEEP
pythonopencvcomputer-visionazure-cognitive-services
Akash Dixit
Akash Dixit

Reputation: 117

How can i pass capture image directly as a binary data for processing in API calling (Microsoft Cognitive Services) using Python

I am using Microsoft cognitive services for face and emotion recognition with python language. Now first i am capturing a image from the webcam using opencv and saving that image in a folder and in API post request i am passing image address for processing then i am getting the desired output. Now i want to save time of processing and want to capture the image from camera and send directly for processing without saving it. How can i do this using python ? Please help me i an newbie in field of programming.

Here's my code: 

while(True):
    ret,img=cam.read()
    faces=faceDetect.detectMultiScale(img,1.3,5)

    for(x,y,w,h) in faces:
        sampleNumber=sampleNumber+1
        cv2.imwrite("dataSet/User."+str(id)+"."+str(sampleNumber)+".jpg",img)
        cv2.rectangle(img,(x,y),(x+w,y+h),(0,255,0),3)
        cv2.waitKey(10)
    cv2.imshow("Face",img)

    img_filename = "C:/Users/Robot 2/Desktop/codes_msc/dataSet/User."+str(id)+"."+str(sampleNumber)+".jpg"
with open(img_filename, 'rb') as f:
    img_data = f.read()
    header = 
    {
    # Request headers for detection
    'Content-Type': 'application/octet-stream',
    'Ocp-Apim-Subscription-Key': subscription_key
     }



  r = requests.post(api_url,
                  params=params,
                  headers=header,
                  data=img_data)
  #Here i don't want to pass img_data as an address i just want to pass image captured

Upvotes: 0

Views: 1361

Answers (1)

cthrash
cthrash

Reputation: 2973

You can encode the image in memory with cv.imencode, and send that to the API. It would look something like the following:

ret,buf = cv.imencode('.jpg', img)

headers = {
    'Content-Type':'application/octet-stream',
    'Ocp-Apim-Subscription-Key':subscription_key }

api_url = 'https://westus.api.cognitive.microsoft.com/face/v1.0/detect'

params = {
    'returnFaceLandmarks':True, 
    'returnFaceAttributes':'emotion,age,gender' }

r = requests.post(api_url,
    params=params,
    headers=headers,
    data=buf.tobytes())

Upvotes: 2

Related Questions