A fundamental problem in object recognition is the development of image representations that are invariant to common transformations such as translation, rotation, and small deformations. There are multiple hypotheses regarding the source of translation invariance in CNNs. One idea is that translation invariance is due to the increasing receptive field size of neurons in successive convolution ...

For strongly directed anisotropic diffusion filtering it is crucial to use numerical schemes with highly accurate directional behaviour. To this end, we introduce a novel algorithm for coherence-enhancing anisotropic diffusion. It applies recently discovered differentiation filters with optimal rotation invariance [10], and comes down to an explicit scheme on a 5×5 stencil. By comparing it with...

Classification of texture images, especially those with different orientation and scale changes, is a challenging and important problem in image analysis and classification. This paper proposes an effective scheme for rotation and scale invariant texture classification using log-polar wavelet signatures. The rotation and scale invariant feature extraction for a given image involves applying a l...

In this paper we proposed a method to design and numerically calculate high-order rotation invariants from Gaussian–Hermite moments. We employed the invariant properties of the Gaussian–Hermite moments discovered earlier in [1] and we showed how to construct rotation Gaussian–Hermite invariants even in the cases when no explicit invariants from geometric moments are available. We verify by expe...

Visual perception of transformation invariance, such as translation, rotation and scaling, is one of the important functions of processing visual information in the Brain. To simulate this perception property, we propose a computational model for perception of transformation. First, we briefly introduce the transformation-invariant basis functions learned from natural scenes using Independent C...

In pattern recognition, it is often necessary to deal with problems to classify a transformed pattern. A neural pattern recognition system which is insensitive to rotation of input pattern by various degrees is proposed. The system consists of a fixed invariance network with many slabs and a trainable multilayered network. The system was used in a rotation-invariant coin recognition problem to ...

Phase correlation is a well-known technique for image registration that is robust to noise and operates in constant time. Scale and rotation invariance can be achieved through means of a log-polar transformation. Unfortunately this method has been historically shown to be unable to handle large rotation or scaling factors, making it unsuitable for many image registration tasks. This paper prese...

The visual illusion of apparent rigid rotation was produced by sequential alternation of two views of the same object in different orientations. The minimum stimulus-onset asynchrony required for the appearance of rigid rotation was a linearly increasing function of the angular difference in orientation between the two views. Variation in the size of the object affected the zero-intercept of th...