CRAFTED_​objectdetector_​GVision_​base64_​public

0 ×

This workflow will detect objects in images, using the Google vision API. The workflow builds upon the results from the workflow 'CRAFTED_get-data-MINT_public:

1. read data about the image from a MongoDB and load the image
2. base65 encode the image
3. execute a request to Google Vision object detection
4. recalculate result to xywh-pixel coordinates
5. write the result in W3C annotation format to a file
6. write the result in W3C annotation format to the MongoDB

Configuration:
- set mongoDB connection to read image metadata [see workflow 'CRAFTED_get-data-MINT_public]
- set CP name and dataset ID
- set API key in Post Request node
- set mongoDB connection to write results
- set path to annotations folder

Requirements:
- MongoDB Compass (see https://www.mongodb.com/products/compass)
- a Google Vision account and API_KEY (see https://cloud.google.com/vision; https://cloud.google.com/vision/docs/object-localizer)

Nodes

• Math Formula4 ×
• JSON Path3 ×
• Missing Value3 ×
• Row Filter3 ×
• String Manipulation3 ×
• Show all 19 nodes

Extensions

• FeatureKNIME Base nodes
• FeatureKNIME Basic File System Connectors
• FeatureKNIME Javasnippet
• FeatureKNIME JSON-Processing
• FeatureKNIME Math Expression (JEP)
• Show all 8 modules

Links

• No links available