# US 2016/0253775 A1 · Methods and systems for designing correlation filter

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2016/0253775 A1
- Document ID: US20160253775A1
- Status: Published application
- Kind: A1
- Application number: 15/055,338
- Filed: 2016-02-26
- Published: 2016-09-01
- Earliest claimed: 2015-02-26
- Assignees: Tata Consultancy Services Limited
- Applicants: Tata Consultancy Services Limited
- Inventors: Amit KALELE; Anubhav Jain; Srinivasa Rao Chalamala; Manoj Karunakaran Nambiar
- Web page: https://patentgrep.com/patent/US20160253775A1

## CPC

- G06T 1/20: Processor architectures; Processor configuration, e.g. pipelining
- G06F 13/4282: {on a serial bus, e.g. I2C bus, SPI bus (on daisy chain buses G06F13/4247)}
- G06T 1/60: Memory management
- G06V 10/7515: {Shifting the patterns to accommodate for positional errors}
- G06V 10/94: Hardware or software architectures specially adapted for image or video understanding
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING

## Abstract

This disclosure relates generally to correlation filters, and more particularly to designing of correlation filter. In one embodiment, a system for designing a correlation filter in a multi-processor system includes a multi-core processor coupled to a first memory and one or more co-processors coupled to one or more respective second memories. The multi-core processor partitions each of a plurality of frames associated with media content into a plurality of pixel-columns, and systematically stores said pixel-columns width-wise in a plurality of temporary matrices by a plurality of threads of the multi-core processor. The plurality of temporary matrices are transferred by the multi-core processor to one or more respective second memories in a plurality of streams simultaneously in an asynchronous mode. A plurality of filter harmonics of the correlation filter are computed by performing compute operations involving at least the plurality of temporary matrices, to obtain the correlation filter.

## Claims

### Claim 1

A computer-implemented method for designing a correlation filter in a multi-processor system having a multi-core processor coupled to a first memory and one or more co-processors coupled to one or more respective second memories, the method comprising:
receiving, by the multi-core processor, a media content comprising a plurality of frames;
partitioning, by the multi-core processor, each frame of the plurality of frames into a plurality of pixel-columns having equal width;
systematically storing the plurality of pixel-columns width-wise in a plurality of temporary matrices, wherein storing the plurality of pixel-columns in the plurality of temporary matrices is performed in parallel by a plurality of threads of the multi-core processor;
transferring, by the multi-core processor, the plurality of temporary matrices to one or more respective second memories in a plurality of streams simultaneously in an asynchronous mode; and
computing, by the one or more co-processors, a plurality of filter harmonics of the correlation filter in the plurality of streams, wherein computing in a stream of the plurality of streams comprises performing compute operations involving at least the plurality of temporary matrices, to obtain the correlation filter.

### Claim 2 (depends on 1)

The method of claim 1, wherein the media content comprises one of a video and a set of images.

### Claim 3 (depends on 1)

The method of claim 1, wherein the multi-core processor comprises a central processing unit (CPU).

### Claim 4 (depends on 3)

The method of claim 3, wherein each of the one or more co-processors of the set of second multi-core processors comprises a graphic processing unit (CPU).

### Claim 5 (depends on 4)

The method of claim 4, wherein the CPU is coupled to the set of GPUs by a bus.

### Claim 6 (depends on 5)

The method of claim 5, wherein the bus is a Peripheral Component Interconnect (PCI) bus.

### Claim 7 (depends on 1)

The method of claim 1, wherein the correlation filter comprises Optimal Trade-Off Circular Harmonic Function (OTCHF) filter

### Claim 8 (depends on 5)

The method of claim 5, wherein systematically storing the plurality of pixel-columns width-wise comprises:
extracting single pixel-column from each of the plurality of frames; and
storing the extracted single pixel-column extracted from each of the plurality of frames in a temporary matrix of the plurality of temporary matrices

### Claim 9 (depends on 1)

The method of claim 1, wherein prior to computing the plurality of filter harmonics of the correlation filter, the method comprises distributing the plurality of temporary matrices in a plurality of sets of batches to each of the plurality of streams, wherein each stream of the plurality of streams comprises a plurality of compute threads operating in parallel.

### Claim 10 (depends on 1)

The method of claim 1, wherein the compute operations comprises linear algebraic and Fast Fourier transform operations.

### Claim 11

A multi-processor system for designing a correlation filter, the system comprising
a first memory;
a multi-core processor coupled to the first memory;
one or more co-processors coupled to the multi-core processor; and
one or more respective second memories associated with the one or more co-processors, wherein the multi-core processor is capable of executing programmed instructions stored in the first memory and the one or more co-processors are capable of executing programmed instructions stored in the one or more respective second memories to:
receive, by the multi-core processor, a media content comprising a plurality of frames;
partition, by the multi-core processor, each frame of the plurality of frames into a plurality of pixel-columns having equal width;
systematically store the plurality of pixel-columns width-wise in a plurality of temporary matrices, wherein storing the plurality of pixel-columns in the plurality of temporary matrices is performed in parallel by a plurality of threads of the multi-core processor;
transfer, by the multi-core processor, the plurality of temporary matrices to one or more respective second memories in a plurality of streams simultaneously in an asynchronous mode; and
compute, by the one or more co-processors, a plurality of filter harmonics of the correlation filter in the plurality of streams, wherein computing in a stream of the plurality of streams comprises performing compute operations involving at least the plurality of temporary matrices, to obtain the correlation filter.

### Claim 12 (depends on 11)

The system of claim 11, wherein the media content comprises one of a video and a set of images.

### Claim 13 (depends on 11)

The system of claim 11, wherein the multi-core processor comprises a central processing unit (CPU).

### Claim 14 (depends on 12)

The system of claim 12, wherein each of the one or more co-processors of the set of second multi-core processors comprises a graphic processing unit (GPU).

### Claim 15 (depends on 13)

The system of claim 13, wherein the CPU is coupled to the set of GPUs by a PCI bus.

### Claim 16 (depends on 11)

The system of claim 11, wherein the correlation filter comprises Optimal Trade-Off Circular Harmonic Function (OTCHF) filter.

### Claim 17 (depends on 11)

The system of claim 11, wherein to systematically store the plurality of pixel-columns width-wise, the multi-core processor is configured by the instructions to:
extract single pixel-column from each of the plurality of frames; and
store the extracted single pixel-column extracted from each of the plurality of frames in a temporary matrix of the plurality of temporary matrices.

### Claim 18 (depends on 11)

The system of claim 11, wherein prior to computing the plurality of filter harmonics of the correlation filter, the multi-core processor is configured by the instructions to distribute the plurality of temporary matrices in a plurality of sets of batches to each of the plurality of streams, wherein each stream of the plurality of streams comprises a plurality of compute threads operating in parallel.

### Claim 19 (depends on 11)

The system of claim 11, wherein the compute operations comprises linear algebraic and fast Fourier transform operations.

### Claim 20

A non-transitory computer-readable medium having embodied thereon a computer program for executing a method for designing a correlation filter in a multi-processor system, the method comprising:
receiving, by the multi-core processor, a media content comprising a plurality of frames;
partitioning, by the multi-core processor, each frame of the plurality of frames into a plurality of pixel-columns having equal width;
systematically storing the plurality of pixel-columns width-wise in a plurality of temporary matrices, wherein storing the plurality of pixel-columns in the plurality of temporary matrices is performed in parallel by a plurality of threads of the multi-core processor;
transferring, by the multi-core processor, the plurality of temporary matrices to one or more respective second memories in a plurality of streams simultaneously in an asynchronous mode; and
computing, by the one or more co-processors, a plurality of filter harmonics of the correlation filter in the plurality of streams, wherein computing in a stream of the plurality of streams comprises performing compute operations involving at least the plurality of temporary matrices, to obtain the correlation filter.
