Authors: Alexandre Alahi, Raphael Ortiz, Pierre Vandergheynst

Keywords:

Keypoint
Binary descriptor
Retina
Sampling
Saccadic
Coarse-to-fine
Orientation

Q1: What is the formula for the retina pattern?

The one difference from BRISK is that the pattern has overlapping circles. In BRISK, they were tangential. Redundancy increases recognition.

The circles are log-polar. In this case, it is similar to Shape Context descriptors, but we do not divide into regions; we create increasingly larger circles on polar lines.

Q2: What do the descriptors contain?

A binary descriptor is a string of bits. A bit corresponds to a pair of receptive fields. If the intensity of the first receptive field is larger than the second, the bit is set to 1; otherwise, it is zero.

Q3: How does the sampling work?

Each circle in the pattern is called a receptive field. A Gaussian kernel is applied to these fields, and their intensities are calculated.

Q4: Is there scale invariance? How?

There is no discussion of scale invariance, but scale invariance seems to arise from the building of the descriptor. Since the circles are created in log-polar orbits and bits are put into the descriptor according to their contribution to recognition, scale invariance follows these.

Q5: How is rotation invariance achieved?

Orientation is calculated using 45 symmetric pairs from the center. It has larger steps than those of BRISK and thus needs lower memory.

Table of Contents