Abstract:
Computer vision requires appropriate image representations. Rotation-invariant feature learning has been proposed using RBM extensions. In this paper, we present an extended novel RBM that explicitly factors for rotation nuisance in 2D image inputs to learn rotation invariant features unsupervised.
While learning invariant features, our model infers an orientation per input image using reconstruction error information. Kullback-Leibler divergence stabilizes and regularizes training.
Our approach learned rotation-invariant features mathematically and experimentally using the _-score. Our method outperforms the current state-of-the-art RBM approaches for rotation invariant feature learning on three benchmark datasets using SVM classifier test accuracy.
Note: Please discuss with our team before submitting this abstract to the college. This Abstract or Synopsis varies based on student project requirements.
Did you like this final year project?
To download this project Code with thesis report and project training... Click Here
