Screengrabs
Enable this model configuration to capture keyframes and slides from a visual meeting/conversation.
Built as a video-first platform, Marsview is the only company in this category that captures keyframes and slides from videos and screen sharing from an online web conference.
Keys | Value |
modelType | screengrabs |
modelConfig | Model Configuration object for screengrabs |
modelConfig | Description | Defaults |
ocr.enable | Boolean to enable or disable OCR | False |
screen_activity.enable | Filter the screengrabs based on the given classes. If this is set to true, then the screengrabs captured by the model will be filtered based on them | True |
screen_activity.classes | A list of classes that the model can recognize. By default all the classes are given. Select the classes that best suit the type of input video. Eg: For meeting, set the class list with values ppt and speaker_presentation | ['drawing', 'ppt', 'screen', 'speaker_face', 'speaker_presentation'] |
Curl
Python
curl --location --request POST 'https://api.marsview.ai/cb/v1/conversation/save_file_link/{{user ID}' \
--header 'Content-Type: application/json' \
--header 'authorization: {{Replace with your authentication token' \
--data-raw '{
"userId":"{{Insert User ID}}",
"txnId": "{{Insert txn ID}}",
"enableModels":[
{
"modelType":"screengrabs",
"modelConfig":{
"ocr":{
"enable":true
},
"screen_activity" : {
"enable" : true,
"classes" : ["drawing","ppt","screen","speaker_face","speaker_presentation"]
}
}
}
]
}'
import requests
user_id = "[email protected]"
auth_token = "replace this with your auth token"
txn_id = "Replace this with your txn id"
request_url = "https://api.marsview.ai/cb/v1/conversation/{user_id}/compute"
def get_screen_grabs():
payload={
"userId":user_id,
"txnId": txn_id,
"enableModels":[
{
"modelType":"screengrabs",
"modelConfig":{
"ocr":{
"enable":true
},
"screen_activity" : {
"enable" : true,
"classes" : ['drawing', 'ppt', 'screen',
'speaker_face', 'speaker_presentation']
}
}
},
]
}
headers = {'authorization': '{}'.format(auth_token)}
response = requests.request("POST", request_url.format(user_id=user_id), headers=headers, json=payload)
print(response.text)
if response.status_code == 200 and response.json()["status"] == "true":
return response.json()["data"]["enableModels"]["state"]["status"]
else:
raise Exception("Custom exception")
if __name__ == "__main__":
get_screen_grabs()
"data": {
"screengrabs": [
{
"shotId": null,
"timeStamp": 13466.666666666666,
"frameNumber": 202,
"meetingActivivty": null,
"title": [],
"text": [],
"confidenceTitle": [],
"confidenceText": [],
"bboxTitle": [],
"bboxText": []
},
]
}
Field | Description |
screenGrabs | List of Screengrab Objects |
shotId | Id of the collection of frames being analyzed |
timeStamp | Start time of the shotId in milliseconds |
frameNumber | Frame number of the frame being analyzed |
meetingActivity | Object specifying the meeting activity, returns null if there is not meeting activity |
title | List of String |
text | List of text identified in the given time frame |
confidenceTitle | List of confidence titles |
confindenceText | List of confidence text |
bboxTitle | Gives a list of 4 parameters with index 0 = X-displacement, 1=Y-displacement, 2=X-end co-ordinate, 3=Y end co-ordinate which represents the coordniates of bounding boxes for the titles identified by the model in the given frame. |
bboxText | Gives a list of 4 parameters with index 0 = X-displacement, 1=Y-displacement, 2=X-end co-ordinate, 3=Y end co-ordinate which represents the coordniates of bounding boxes for the text areas identified by the model in the given frame. |
Last modified 2yr ago