Filter bank based phase correlation architecture for motion estimation

1. A system for identifying motion between a previous frame and a current frame, the system comprising:

a fast Fourier transform calculator configured to generate corresponding low pass frequency domain representations of previous frame data and current frame data and corresponding high pass frequency domain representations of the previous frame data and the current frame data;

a phase difference calculator configured to calculate a first phase difference between the previous frame data low pass frequency domain representation and the current frame data low pass frequency domain representation, the phase difference calculator further configured to calculate a second phase difference between the previous frame data high pass frequency domain representation and the current frame data high pass frequency domain representation;

an inverse Fourier transform calculator configured to generate a first inverse Fourier result based on the first phase difference and a second inverse Fourier result based on the second phase difference; and

a motion vector calculator configured to generate two motion vectors based on the first inverse Fourier result and the second inverse Fourier result;

wherein the generated two motion vectors are used to generate a display on a display device.

2. The system of claim 1 , further comprising a decomposition filter that receives the previous frame data and the current frame data, the decomposition filter comprising a high pass filter and a low pass filter, the decomposition filter configured to provide a high pass filter output and a low pass filter output to the fast Fourier transform calculator.

3. The system of claim 1 , wherein the first inverse Fourier result is a first phase plane correlation surface, and wherein the second inverse Fourier result is a second phase plane correlation surface.

4. The system for claim 3 , wherein the motion vector calculator comprises a peak searcher that identifies a first peak on the first phase plane correlation surface and a second peak on the second phase plane correlation surface.

5. The system of claim 4 , wherein the two motion vectors correspond to a location of the first identified peak and a location of the second identified peak.

6. The system of claim 5 , wherein the locations of the first identified peak and the second identified peak correspond to a movement in a video from the previous frame to the current frame.

7. The system of claim 4 , wherein the peak searcher identifies a first plurality of additional peaks on the first phase plane correlation surface and a second plurality of additional peaks on the second phase plane correlation surface.

8. The system of claim 2 , wherein the previous frame data and the current frame data correspond to frame-blocks 64×32 in size, and wherein the low pass filter is a multi-tap filter having a block size of 15×7.

9. The system of claim 2 , wherein the high pass filter comprises a gradient edge based detector.

10. The system of claim 9 , wherein the gradient edge based detector is a Sobel edge detector.

11. The system of claim 1 , wherein the fast Fourier transform calculator utilizes a radix 2 decimation-in-time implementation.

12. The system of claim 1 , wherein the fast Fourier transform calculator has an input resolution of 2×16 bits and an output resolution of 2×16 bits.

13. The system of claim 1 , wherein the phase difference calculator uses a CORDIC implementation, wherein the first phase difference and the second phase difference are represented with 8 bit resolution, and the phase difference calculator has a latency of 6 clock cycles.

14. The system of claim 1 , wherein the phase difference calculator comprises a phase subtractor and a lookup table.

15. A method of identifying motion between a previous frame and a current frame, the method comprising:

decomposing received previous frame data and current frame data into a high pass component and a low pass component;

performing a fast Fourier transform of the high pass components and low pass components of the previous frame data and the current frame data;

calculating a first phase difference between the high pass component of the current frame data and the high pass component of the previous frame data;

calculating a second phase difference between the low pass component of the current frame data and the low pass component of the previous frame data;

performing an inverse fast Fourier transform on the first phase difference and the second phase difference to generate a first inverse Fourier result and a second inverse Fourier result; and

calculating a first motion vector and a second motion vector based on the first inverse Fourier result and the second inverse Fourier result;

wherein the generated two motion vectors are used to generate a display on a display device.

16. The method of claim 15 , wherein the first inverse Fourier result is a first phase plane correlation surface, and the second inverse Fourier result is a second phase plane correlation surface.

17. The method of claim 16 , wherein calculating a first motion vector and a second motion vector further comprises searching for a first peak of the first phase plane correlation surface and a second peak of the second phase plane correlation surface.

18. The method of claim 17 , wherein the first motion vector and the second motion vector are calculated based on locations of the first peak on the first phase plane correlation surface and the second peak of the second phase plane correlation surface.

19. The method of claim 18 , wherein searching for peaks comprises identifying a maximum value in a 3×3 neighborhood, and zeroing all non-maximum values in the 3×3 neighborhood.

20. The method of claim 18 , wherein searching for peaks further comprises identifying a first plurality of peaks from the first phase plane correlation surface and a second plurality of peaks from the second phase plane correlation surface.