'Train an NN using SIFT information
I have keypoint information using the SIFT descriptor for an image dataset/s. I need to train a network to predict the characteristic scales (labels) of the keypoints. My question is regarding the input data to the network. How can I feed the input image and the keypoint information (datapoint) with the corresponding label into the network? Can I make use of the coordinates of the keypoint as an anchor point in an image? and if yes, how?
Sources
This article follows the attribution requirements of Stack Overflow and is licensed under CC BY-SA 3.0.
Source: Stack Overflow
| Solution | Source |
|---|