Rotation Invariance Neural Network
This addresses the need for robust image recognition in tasks like symbol detection, though it appears incremental as it builds on existing CNN architectures.
The paper tackles the problem of achieving rotation invariance in 2-D symbol recognition by introducing a new cyclic convolutional layer in CNNs, enabling detection of position and orientation for multiple non-overlapping targets and one-shot learning in some cases.
Rotation invariance and translation invariance have great values in image recognition tasks. In this paper, we bring a new architecture in convolutional neural network (CNN) named cyclic convolutional layer to achieve rotation invariance in 2-D symbol recognition. We can also get the position and orientation of the 2-D symbol by the network to achieve detection purpose for multiple non-overlap target. Last but not least, this architecture can achieve one-shot learning in some cases using those invariance.