1 Answer
- Newest
- Most votes
- Most comments
1
To match similar non-face images using Amazon Rekognition and Python, focus on the DetectLabels function. This API action identifies objects, scenes, and concepts in images.
How to :
- Set up your AWS SDK and create a Rekognition client in Python.
- Use DetectLabels to analyze each image, obtaining labels that describe the image's contents.
- Filter the images by reviewing these labels and their confidence scores, keeping those that match your criteria (e.g., forest scenes but not beach scenes).
Relevant content
- asked 6 months ago
- AWS OFFICIALUpdated 2 years ago
- AWS OFFICIALUpdated 2 years ago
- AWS OFFICIALUpdated 2 years ago
I appreciate the quick answer. This seems to be a very good way to great granular information regarding an image. If you have a scenario with a 'beach' photo properly handling labels and you want to search for 'like' photos in a pile of, say 5,000, I assume step one will be to create a db that captures what Rekog finds for all 5,000 images and then when doing the 'like this' zip through that db looking for the labels as you described. Should be fun to code up. Any other thoughts, let me know. Thanks again!