Abstract: A method (200) of enhanced peak detection in a correlation signal (g(x)) is disclosed. The method (200) comprises identifying (215) in the correlation signal (g(x)) positions where the modulus value of the correlation signal is above a first predetermined value. Each of the identified positions is then processed by forming (220) a sub-signal from the correlation signal by isolating a region of samples, the region comprising substantially equal numbers of samples on opposite sides of the identified position. The sub-signal is then up-sampled (221, 222, 225) to a higher resolution using Fourier interpolation. The up-sampled sub-signal is searched (230) for a peak value, and when the peak value is greater than a second predetermined value, parameters associated with the peak value is stored (236).

Abstract: A system (100) and a method (500) are described for determining a two-dimensional position of a location in an image. The method (500) starts by imaging (510) a two-dimensional pattern (440). The two-dimensional pattern comprises a plurality of at least partially overlapping two-dimensional sub-patterns (410, 420, 430). The sub-patterns (410, 420, 430) repeat with different spatial periods to form the two-dimensional pattern, and the spatial period of the sub-patterns are anharmonic. A two-dimensional offset for each of the sub-patterns is then determined (540) at the location in the image formed by the imaging. The two-dimensional position is determined from said two-dimensional offsets.

Abstract: A method (200) is disclosed of detecting one or more patterns embedded in an image. Each pattern embedded in the image has been formed from a one-dimensional basis function. The method (200) starts by calculating (210) a projective transform of the image. A 1-D correlation is then calculated (220) between the projective transform and the basis function for a selection of angles. Finally, one or more peaks of the correlation are found (230). The position of each of the peaks provides spatial parameters of one of the one or more embedded patterns.

Abstract: A local phase filter (315) for pre-processing data (310) including a watermark (320) is disclosed. The data (310) is pre-processed before detecting the watermark by correlation of the data (310) with the watermark (320). The local phase filter (315) starts by dividing the data (310) into a set of smaller sub-spaces. The frequency coefficients of the data contained in each sub-space are then computed, and the frequency amplitudes of the frequency coefficients are then spectrally shaped. Preferably, the amplitudes of the frequency coefficients of the sub-space are set to a predetermined function. Finally, filtered watermarked data is formed, having the same size as that of the data (310), by adding the filtered data of each sub-space together in the sub-space's original position.

Abstract: A method (750) of encoding a value is disclosed. In one implementation the value is encoded into an image (700). The method (750) operates by defining (403) a first ordered set of positions; determining (405) a number of marks for encoding the value in the ordered set of positions; determining (409) a selection of the first ordered set of positions using combinatorial theory with the number of marks and the number of positions in the first ordered set of positions for encoding the value; and placing (411) marks at the selection of the first ordered set of positions. Preferably the value is encoded by encoding a base value and an offset value, the base value is encoded by the number of marks, and the offset value is encoded by the selection of positions. Preferably the marks are basis patterns forming a watermark (710).

Abstract: Methods (450, 750) are disclosed for embedding a watermark into an image (400, 700). The watermark comprises at least one basis pattern. A real part of the basis pattern(s) (410, 710) is added to the image (400, 700) to form a watermarked image (420, 730). The basis pattern(s) has scale and rotation invariant properties. The pattern(s) is preferably added to the image at a low intensity to make the pattern(s) invisible or imperceptible to the human visual system under normal viewing conditions. Methods (800, 900) are also disclosed for detecting a watermark from a watermarked image (810, 910). The watermark image (810, 910) is correlated with the basis pattern to provide a result image (830, 950).

Abstract: A method (200) is disclosed of detecting one or more patterns embedded in an image. Each pattern embedded in the image has been formed from a one-dimensional basis function. The method (200) starts by calculating (210) a projective transform of the image. A 1-D correlation is then calculated (220) between the projective transform and the basis function for a selection of angles. Finally, one or more peaks of the correlation are found (230). The position of each of the peaks provides spatial parameters of one of the one or more embedded patterns.

Abstract: A local phase filter (315) for pre-processing data (310) including a watermark (320) is disclosed. The data (310) is pre-processed before detecting the watermark by correlation of the data (310) with the watermark (320). The local phase filter (315) starts by dividing the data (310) into a set of smaller sub-spaces. The frequency coefficients of the data contained in each sub-space are then computed, and the frequency amplitudes of the frequency coefficients are then spectrally shaped. Preferably, the amplitudes of the frequency coefficients of the sub-space are set to a predetermined function. Finally, filtered watermarked data is formed, having the same size as that of the data (310), by adding the filtered data of each sub-space together in the sub-space's original position.

Abstract: A method (750) of encoding a value is disclosed. In one implementation the value is encoded into an image (700). The method (750) operates by defining (403) a first ordered set of positions; determining (405) a number of marks for encoding the value in the ordered set of positions; determining (409) a selection of the first ordered set of positions using combinatorial theory with the number of marks and the number of positions in the first ordered set of positions for encoding the value; and placing (411) marks at the selection of the first ordered set of positions. Preferably the value is encoded by encoding a base value and an offset value, the base value is encoded by the number of marks, and the offset value is encoded by the selection of positions. Preferably the marks are basis patterns forming a watermark (710).