KR101221227B1 - Method and apparatus for estimating motion using region allocation based on multi core - Google Patents

Abstract

PURPOSE: A motion estimating apparatus using a multi-core based region allocation is provided to lower the complexity of a spiral shape searching method by interactively allocating the position of each object searcher and excluding a similar operation process in advance. CONSTITUTION: A motion estimation apparatus produces at least one object searcher searching a reference block in a previous video frame which is the same object with a current block in the current image frame of received image data(S200). The apparatus divides each domain in the previous image frame. The apparatus allocates the object searcher in the allocated domain(S300). The apparatus produces a motion vector of which a position change between the reference block and the current block becomes a minimum value by the object search of the object searcher in the previous image frame. The apparatus determines a reference block in the previous image frame(S400). [Reference numerals] (AA) Start; (BB) End; (S100) Receiving image data from a plurality of frames; (S200) Producing at least one object searcher for searching an object same as an object which is located in a current block in a current image frame of the image data, within a previous video frame; (S300) Allocating the object searcher by dividing each area within the previous image frame; (S400) Producing a motion vector of which an object position change becomes a minimum value by the object search of the object searcher in the previous image frame, and determining a reference block in the previous image frame

Description

METHOD AND APPARATUS FOR ESTIMATING MOTION USING REGION ALLOCATION BASED ON MULTI CORE}

Embodiments of the present invention relate to a motion estimation method and apparatus using multi-core based region allocation, and more particularly, to a motion estimation method using multi-core based region allocation capable of minimizing computational complexity by segmenting and allocating image regions. Relates to a device.

In general, video data has a lot of overlapping information, and therefore requires a large amount of data to process a digital video. Recently, various transmission and storage media using VLSL and signal processing technologies have appeared to process, store, and transmit such digital images. However, considering the bandwidth of the communication line required for transmission, the storage capacity required for information storage, and the like, a data compression technique having a high compression ratio is required.

Among the methods proposed to meet this demand, there is an image motion estimation method. The method finds the block most similar to the block of the current picture in the reference picture or reference pictures, and represents the displacement between the block of the current picture and the corresponding block in the reference picture as a motion vector.

Among motion estimation methods of a video, there is a full search method. This is a method of first determining a search region in a reference image and determining the displacement between the current block and the block having the smallest prediction error among all possible blocks within the maximum displacement of the search region as a motion vector. The full-area search method can obtain motion information with very high accuracy when the maximum motion does not leave the search area.

However, a high memory bandwidth is required to implement a motion estimation and compensation algorithm used for video compression and decompression. In a real time encoder, a search area for motion estimation is limited by limited resources.

Therefore, a motion estimation method that can have a high compression rate without limiting the search area has been required, and Korean Patent Publication No. 10-0492127 discloses such a technique.

However, the process of finding a reference block using such a motion estimation method has a problem in that the computation amount and complexity are high, resulting in an increase in time caused by an operation and an accompanying system construction cost.

Korean Registered Publication No. 10-0492127

The embodiment of the present invention was devised to solve the above-described problem, and by interleaving the position of each object searcher to exclude similar computational processes, thereby reducing the complexity of the spiral search method. An object of the present invention is to provide a motion estimation method and apparatus using the base area allocation.

In an embodiment of the present invention, since the number of cores of each object searcher is provided and matched to find the reference block quickly, and the multi-cores are sequentially or simultaneously driven to find the reference block for each region, It is an object of the present invention to provide a motion estimation method and apparatus using multi-core based region allocation that can reduce the complexity of computation and the time.

An embodiment of the present invention provides a motion estimation method and apparatus capable of minimizing a system construction cost and time required for encoding, decoding, and compressing an image by minimizing the complexity of motion estimation using an object searcher and multiple cores. It aims to do it.

In order to achieve the object as described above, the present invention, in the embodiments of the present invention, the step of receiving the image data consisting of a plurality of frames, the same object located in the current block in the current image frame of the image data, the previous image frame Generating at least one object searcher to search within, dividing and allocating the object searcher in each region in the previous image frame, and performing the object search by the object searcher in the previous image frame The problem can be solved using a motion estimation method using multi-core based region allocation, which includes calculating a motion vector having a minimum position movement and determining a reference block in a previous image frame.

In one or more embodiments, a receiver configured to receive image data including a plurality of frames, and generate at least one object searcher to search for an object identical to an object located in a current block within a current image frame of the image data within a previous image frame. A generation unit for allocating the object searcher by dividing each area in the previous image frame, and a motion vector in which the object position is minimized by performing the object search by the object searcher in the previous image frame. It is possible to solve the problem by using a motion estimation apparatus using multi-core-based region allocation that includes a determination unit for calculating a vector and determining a reference block in a previous image frame.

As described above, the disclosed technology of the present invention having the above-described configuration is capable of lowering the complexity of the spiral search method by interleaving the position of each object searcher and excluding similar calculation processes. A motion estimation method and apparatus using core-based region allocation can be provided.

In addition, the present invention minimizes the complexity of motion estimation by using an object searcher and multicore, and thus, motion estimation using multi-core based region allocation, which minimizes the system construction cost and time required for encoding, decoding, and compressing an image. A method and apparatus can be provided.

1 is a conceptual diagram illustrating a motion estimation according to an embodiment of the present invention.
2 is a conceptual diagram illustrating a motion estimation according to an embodiment of the present invention.
3 is a diagram illustrating a spiral search method according to an embodiment of the present invention.
4 is a block diagram illustrating a motion estimation apparatus according to an embodiment of the present invention.
5 is a view for explaining a motion estimation method according to an embodiment of the present invention.
6 is a diagram illustrating a motion vector calculated according to an embodiment of the present invention.
7 is a flowchart illustrating a motion estimation method according to an embodiment of the present invention.
8 is a view for explaining a motion estimation method according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the following description, a detailed description of known techniques well known to those skilled in the art may be omitted.

In describing the constituent elements of the present invention, the same reference numerals may be given to constituent elements having the same name, and the same reference numerals may be given thereto even though they are different from each other. However, even in such a case, it does not mean that the corresponding component has different functions according to the embodiment, or does not mean that the different components have the same function. It should be judged based on the description of each component in the example.

In the following description of the embodiments of the present invention, a detailed description of known functions and configurations incorporated herein will be omitted when it may make the subject matter of the present invention rather unclear.

In addition, in describing the component of this invention, terms, such as 1st, 2nd, A, B, (a), (b), can be used. These terms are intended to distinguish the constituent elements from other constituent elements, and the terms do not limit the nature, order or order of the constituent elements. When a component is described as being "connected", "coupled", or "connected" to another component, the component may be directly connected or connected to the other component, Quot; may be "connected," "coupled," or "connected. &Quot;

It should be understood that the singular " include "or" have "are to be construed as including a stated feature, number, step, operation, component, It is to be understood that the combination is intended to specify that it does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, or combinations thereof.

In each step, the identification code (e.g., a, b, c, etc.) is used for convenience of explanation, the identification code does not describe the order of each step, It may occur differently from the order specified. That is, each step may occur in the same order as described, may be performed substantially concurrently, or may be performed in reverse order.

Hereinafter, embodiments according to the present invention will be described in detail with reference to the accompanying drawings.

1 is a conceptual diagram illustrating a motion estimation according to an embodiment of the present invention, FIG. 2 is a conceptual diagram illustrating a motion estimation according to an embodiment of the present invention, and FIG. 3 is a spiral search according to an embodiment of the present invention. A diagram illustrating the method.

First, in the video field, when a current block is compressed, a reference block is found in a previous frame image and the difference component is compressed and transmitted. If the size of the reference block and the current block is MxN, the reference block is determined as a position that minimizes the difference between the current block as shown in Equation 1 below.

Where Ic is the current image frame, Ir is the previous image frame, Ic (x, y) is the x, y position pixel value of the Ic frame, and Ir (x, y) is the x, y position pixel value of the Ir frame. Where dx and dy are values indicating the position of the reference picture, and (dx, dy) is referred to as a motion vector.

That is, (a) assuming that there is a current block of size MxN existing in the t th frame, (b) the difference between the current block and the sum absorptive difference (SAD) in the wxw search window around the current block in the t-1 th frame. Finds the block with the smallest position, it can be set as a reference block.

Then, (c) the position change between the current block and the reference block is recognized as a motion vector. The process of finding the reference block is called motion estimation (ME), and since the complexity of the motion estimation is very high, various algorithms have been developed.

The present invention proposes a method for improving spiral search, which is one of motion estimation methods. To this end, the present invention effectively partitions an area for finding a reference block using a multicore system.

 Referring to FIG. 3, the spiral search method takes an origin as a reference point in a previous image frame and searches for a reference block most similar to the center from the center. The helical search order is as shown in Equation 2 below, and the helical search method is calculated by substituting the distance data D of two blocks, the current block and the reference block at each position, into Equation 1 above.

In this case, as soon as the reference block, that is, the position where the distance data is minimized, is found, the subsequent search position may be omitted or simplified. That is, as soon as the distance data D detects a minimum or small location, the complexity of the spiral search method may be lowered.

4 is a block diagram for explaining a motion estimation apparatus according to an embodiment of the present invention, FIG. 5 is a diagram for explaining a motion estimation method according to an embodiment of the present invention, and FIG. 6 is an embodiment of the present invention. It is a figure which shows the motion vector computed accordingly.

4 and 5 are diagrams for explaining a motion estimation apparatus according to an embodiment of the present invention. Since the receiver 10, the generator 30, the allocator 50, the determiner 70, and the division and the allocated region illustrated in FIG. 5 are merely illustrated for convenience of description, The motion estimation apparatus 1 of the present application is not to be classified and set in FIG. 4 and is not limited to those shown in FIGS. 4 and 5.

Referring to FIG. 4, the receiver 10 may receive image data composed of a plurality of frames. This may be for any one of encoding, decoding, encoding, and decoding video or video data. Operations other than the configuration described in the present invention are obvious to those skilled in the art related to the encoding, decoding, encoding, and decoding techniques, and thus detailed description thereof will be omitted.

The generation unit 30 may generate at least one object searcher that searches for a reference block in a previous image frame that is the same object as the current block in the current image frame of the image data. Here, the object searcher may match 1: 1 with at least one core. That is, the object searcher is a configuration for searching for a reference block that is the same object in the previous image frame with reference to the current block of the current image frame. At least one core may be provided to drive the configuration. Here, the object searcher itself may be the core, or the core may be the object searcher itself, and may be variously configured according to a system and an implementation method.

In addition, the object searcher may use a spiral search algorithm, and at least one object searcher may search for a reference block that is an object simultaneously or sequentially. For example, as shown in FIG. 5, the search may be started or sequentially searched to find the same reference block as the current block of the current image frame while being spirally assigned to each divided area as a reference point. have. Alternatively, when the cores for driving the object searcher are matched 1: 1, each core may start driving for searching simultaneously to find the same reference block as the current block of the current image frame, or sequentially driving for searching. You can also do

4 and 5, the allocator 50 may allocate an object searcher by dividing each area of the previous image frame. The allocator 50 may allocate the object searcher interleaved in the previous image frame. In other words, rather than dividing the region continuously, the search position may be divided to be allocated to each object searcher by interleaving.

The reason for this is that distance data is similarly calculated in the neighboring area when performing the motion estimation process, and it is possible to reduce the complexity by performing interleaving rather than performing the motion estimation of the specific area for each object searcher. It may be divided and allocated as shown in Equation 3 below.

Where n is the n-th partition.
The determiner 70 may search for a reference block that is the object of the object searcher in the previous image frame, calculate a motion vector in which the position movement between the current blocks is the minimum value, and determine the reference block in the previous image frame.

The present invention can be effectively used in a digital video compression device such as a digital video recorder, a camcorder, a broadcast equipment. It can also be applied to digital video recording devices such as real-time digital broadcasting equipment, personal digital video recorders, and digital camcorders. In particular, it may be more effective when there is a relatively limited compression capacity and device driving resources, such as a personal portable device.

7 is a flowchart illustrating a motion estimation method according to an embodiment of the present invention.

First, image data composed of a plurality of frames is received (S100).

Then, at least one object searcher is generated to search for a reference block in a previous image frame that is the same object as the current block in the current image frame of the image data (S200).

The generated object searcher is divided and allocated to each region in the previous image frame (S300), and the object searcher searches for the reference block in the previous image frame to calculate a motion vector in which the position movement between the current blocks is the minimum value. And determining a reference block in the previous image frame (S400).

8 is a view for explaining a motion estimation method according to an embodiment of the present invention.

The H.264 / AVC encoder uses a Lagrangian optimization technique to find the optimized motion vector in the inter-coded block. High performance can be achieved by providing optimized bit allocation for motion vectors.

For example, there may be a number of ways in which the spiral search technique may be parallelized, such as a search window that may be separated into different scan patterns. For example, a simple method of increasing the driving length by using the sub window continuously may be used until the search window is completed as shown in (a).

However, the reduction in complexity is a matter of how fast the encoder finds pixel positions with low bit rate distortion costs. The bit rate distortion costs of the proximal positions are strongly linked to each other. Thus, if the search window is divided in succession, and each thread has a value independently in the divided region, each thread can be responsible for the region having a small bit rate distortion cost.

This result leads to the conclusion that different threads must evaluate candidate positions because the current minimum bit rate distortion cost is high. Accordingly, a motion estimation method according to an embodiment of the present invention capable of parallelizing spiral search in a multithreaded hardware structure is proposed.

In the motion estimation method, a number of threads that an encoder can utilize is referred to as N. Then, assuming that the motion search window is divided into N distributed regions as Ln, Ln is expressed by Equation 4 below.

Here, n is 0 or more and N or less, and% means a modulus operator. This method according to an embodiment of the present invention assigns each thread to a W / N pixel position, and the thread evaluates the assigned pixel position independently. Referring to (b), an example of division by the method of Equation 6 is shown. The interleaved allocation of the present invention may be that each thread is arranged as shown in (b).

When starting the motion search method, the master thread creates a number of threads. Multiple threads are assigned to their search areas based on Equation 4 above, and each thread can search for an optimal location within its search area.

The processing time of a thread may vary as the number of search areas varies. After completing the parallel motion estimation method, multiple threads can be synchronized at the end of the search process. Only the main thread can continue processing.

The present invention is not necessarily limited to these embodiments, as all the constituent elements constituting the embodiment of the present invention are described as being combined or operated in one operation. In other words, within the scope of the present invention, all of the components may be selectively operated in combination with one or more. In addition, although all of the components may be implemented as one independent hardware, some or all of the components may be selectively combined to perform a part or all of the functions in one or a plurality of hardware. As shown in FIG. In addition, such a computer program may be stored in a computer-readable medium such as a USB memory, a CD disk, a flash memory, etc., and read and executed by a computer, thereby implementing embodiments of the present invention. As the storage medium of the computer program, a magnetic recording medium, an optical recording medium, a carrier wave medium, or the like may be included.

Furthermore, all terms including technical or scientific terms have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs, unless otherwise defined in the Detailed Description. Terms used generally, such as terms defined in a dictionary, should be interpreted to coincide with the contextual meaning of the related art, and shall not be interpreted in an ideal or excessively formal sense unless explicitly defined in the present invention.

The above description is merely illustrative of the technical idea of the present invention, and those skilled in the art to which the present invention pertains may make various modifications and changes without departing from the essential characteristics of the present invention. In addition, the embodiments disclosed in the present invention are not intended to limit the scope of the present invention but to limit the scope of the technical idea of the present invention. Accordingly, the scope of protection of the present invention should be construed according to the claims, and all technical ideas within the scope of equivalents should be interpreted as being included in the scope of the present invention.

1: motion estimation device
10: receiver 30: generator
50: assigning unit 70: determining unit

Claims (13)

A motion estimation method performed in a motion estimation apparatus using multicore based region allocation,
Receiving image data consisting of a plurality of frames;
Generating at least one object searcher for searching for a reference block in a previous image frame that is the same object as the current block in the current image frame of the image data;
Allocating the object searcher by dividing each area in the previous image frame; and
Determining, by the object searcher, a reference block in the previous image frame to calculate a motion vector having a minimum change in position with the current block and determining a reference block in the previous image frame
Motion estimation method using multi-core based region allocation comprising a.
The method of claim 1,
And the object searcher is interleaved in the previous image frame.
The method of claim 1,
The object searcher is a motion estimation method using multi-core based region allocation, characterized by using a spiral search algorithm
The method of claim 1,
Location of the object searcher is located as shown in the following equation.

Where n is the nth partition
The method of claim 4, wherein
And the at least one object searcher searches for reference blocks that are objects simultaneously or sequentially.
The method of claim 1,
And the at least one object searcher is 1: 1 matched to at least one core.
In the motion estimation apparatus using multi-core based region allocation,
Receiving unit for receiving the image data consisting of a plurality of frames,
A generation unit generating at least one object searcher for searching for a reference block in a previous image frame that is the same object as the current block in the current image frame of the image data;
An allocator for dividing and assigning the object searcher to each region in the previous image frame;
A determination unit configured to search for a reference block that is an object in the previous image frame, calculate a motion vector in which the positional movement of the reference block that is an object becomes a minimum value, and determine a reference block in the previous image frame
Motion estimation apparatus using multi-core based region allocation comprising a.
The method of claim 7, wherein
And the allocator interleaves the object searcher within the previous image frame.
The method of claim 7, wherein
The object searcher is a motion estimation apparatus using multi-core based region allocation, characterized by using a spiral search algorithm
The method of claim 7, wherein
The location of the object searcher is located as shown in the equation below.

Where n is the nth partition
11. The method of claim 10,
And the at least one object searcher searches for reference blocks that are objects simultaneously or sequentially.
The method of claim 11,
And the at least one object searcher is 1: 1 matched with at least one core.
A computer-readable recording medium in which a program for executing the method of any one of claims 1 to 6 is recorded.

Images