KR20120124782A - Apparatus and method for processing digital image - Google Patents

Abstract

PURPOSE: A digital image processing apparatus and a method thereof are provided to increase recognition ratio according to brightness by increasing recognition ratio of a neural skin color by using normalized skin color data. CONSTITUTION: A DSP(Digital Signal Processor) converts an RGB(Red, Green, Blue) signal of an image, in which motion is detected, into a YCbCr signal(503). The DSP normalizes a color signal(505). The DSP detects a skin by a back propagation algorithm(507). The DSP calculates a location of the detected pixel(509). The DSP operates zoom based on the calculated location(511). [Reference numerals] (501) Detecting movement; (503) Color change(R/G/B→YCbCr); (505) Normalization Dr=(Cr+n/Y)*m Db=(Cb+n/Y)*m; (507) Detecting a skin by a back propagation algorithm; (509) Calculating a location of the detected pixel; (511) Operating zoom based on the calculated location; (AA) Start; (BB) End

Description

Apparatus and method for processing digital image}

The present invention relates to a digital image processing apparatus and method for performing image processing using an image sensor.

When motion occurred in a general surveillance system, the camera zoomed according to the movement. In particular, when a movement occurs, the camera zoom is used to zoom in on a moving object to monitor the object in more detail.

Conventional surveillance systems are often smart zoom according to the motion in consideration of the movement, there was a problem that the movement of the object by other influences such as climatic influences (wind, rain) also works the same. In particular, there are many cases where a person, especially a face, is to be mainly monitored among the moving objects. In this case, the smart zoom is easily operated even when the object is moved. In addition, the conventional face recognition technology takes a long time and has a large execution time, which makes it difficult to embed an algorithm into a small surveillance camera system. Therefore, there is a need for a simple technology of monitoring only the face area of a person in real time when a motion occurs.

The technical problem to be solved by the present invention is to provide a digital image processing apparatus and method for performing a zoom using the normalized skin color data in order to increase the recognition rate according to the illumination.

A digital image processing method according to an embodiment of the present invention includes the steps of (a) detecting the movement of the input image; (b) normalizing a color signal of the detected motion image by using the luminance signal of the detected motion image; (c) detecting skin from the normalized signal using a backpropagation algorithm; And (d) zooming around the area detected by the skin.

In the present invention, after normalizing in step (b), a value obtained by adding a first normalization factor to the color signal may be divided by the luminance signal, and then a normal signal may be output by multiplying a second normalization factor.

In the present invention, the first normalization factor may be a bit shift coefficient for making the color signal positive, and the second normalization factor may be a value between 0-1 for generating the backpropagation algorithm input value. .

According to another aspect of the present invention, there is provided a digital image processing apparatus including a motion detector configured to detect a motion by using an input image signal and a previous image signal; A normalizer for normalizing a color signal of the detected motion image by using the luminance signal of the detected motion image; A skin detector for detecting skin from the normalized signal using a backpropagation algorithm; And a zoom controller configured to output a zoom performing signal centering on the area detected by the skin.

In the present invention, the normalization unit may divide a value obtained by adding a first normalization factor to the color signal by the luminance signal, and then multiply a second normalization factor to output a normal signal.

In the present invention, the first normalization factor may be a bit shift coefficient for making the color signal positive, and the second normalization factor may be a value between 0-1 for generating the backpropagation algorithm input value. .

As described above, according to the present invention, by increasing the recognition rate of the neural network skin color by using the normalized skin color data, the recognition rate according to the illumination can be increased, and the calculation amount and the learning time can be reduced, so that a quick zoom process is possible. .

1 is a diagram illustrating a surveillance system to which a digital image processing apparatus is applied according to an exemplary embodiment.
2 is a block diagram illustrating a configuration of a digital image processing apparatus according to an exemplary embodiment.
3 is a detailed block diagram of the digital signal processor of FIG. 2.
FIG. 4 is a diagram illustrating weights generated through pre-learning by the back-propagation algorithm of FIG. 3.
5 is a flowchart illustrating an operation of a digital image processing method according to another embodiment of the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS The present invention is capable of various modifications and various embodiments, and specific embodiments are illustrated in the drawings and described in detail in the detailed description. However, this is not intended to limit the present invention to specific embodiments, it should be understood to include all transformations, equivalents, and substitutes included in the spirit and scope of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

The terms first, second, etc. may be used to describe various elements, but the elements should not be limited by terms. Terms are used only for the purpose of distinguishing one component from another.

The terminology used in this application is used only to describe a specific embodiment and is not intended to limit the invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this application, the terms "comprise" or "have" are intended to indicate that there is a feature, number, step, operation, component, part, or combination thereof described in the specification, and one or more other features. It is to be understood that the present invention does not exclude the possibility of the presence or the addition of numbers, steps, operations, components, components, or a combination thereof.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. Referring to the accompanying drawings, the same or corresponding components are denoted by the same reference numerals, do.

1 is a diagram illustrating a surveillance system to which a digital image processing apparatus is applied according to an exemplary embodiment.

Referring to FIG. 1, the surveillance cameras 101, 111, 121 communicate with a host device 2 as a host device by means of communication signals Dcom, while a video signal Svid of a live view is present. Is transmitted to the host device 2. The video signal Svid received at the host device 2 is displayed via a display device, while being stored in a recording device, for example a hard disk drive.

Each of the surveillance cameras 101, 111, 121 transmits a video signal Svid to the host device 2 while communicating with the host device 2 via a coaxial cable. Accordingly, the communication signals Dcom are transmitted and received in the vertical blank interval of the video signal Svid transmitted to the host device 2.

In response to control signals from the host device 2, the surveillance cameras 101, 111, 121 each perform panning of left and right rotations and tilting of up and down rotations.

Here, the host device 2 controls the operations of each of the surveillance cameras 101, 111, and 121 according to the surveillance camera control method according to an embodiment of the present disclosure. Related contents will be described in detail with reference to FIGS. 2 to 4.

2 is a block diagram illustrating a configuration of a digital image processing apparatus according to an exemplary embodiment.

2, the digital image processing apparatus includes an optical system (OPS), a photoelectric conversion unit (OEC), a correlation double sampler and analog-to-digital converter (CDS-ADC) 210, a timing circuit 220, and a digital as a controller. Digital signal processor (DSP) 230, video-signal generator 240, aperture motor (M _A ), zoom motor (M _Z ), focus motor (M _F ), filter motor (M _D ), panning The motor M _P , the tilting motor M _T , the driver 250, the communication interface 260, the micro-computer 270, and the lighting unit 280 are included.

The optical system OPS including the lens unit and the filter unit optically processes light from a subject.

The lens unit of the optical system OPS includes a zoom lens and a focus lens. In the filter section of the optical system OPS, an optical low pass filter (OLPF) used in a night mode of operation removes high frequency optical noise. Infra-red cut filters (IRFs) used in daytime mode of operation block the infrared component of incident light.

A photoelectric converter (OEC) of a charge coupled device (CCD) or a complementary metal-oxide-semiconductor (CMOS) converts light from an optical system (OPS) into an electrical analog signal. Here, the digital signal processor 230 controls the timing circuit 220 to control operations of the photoelectric converter OEC and the correlation double sampler and analog-to-digital converter 210.

The CDS-ADC 210 processes the analog video signal from the photoelectric converter (OEC), removes the high frequency noise, adjusts the amplitude, and converts the digital video data. This digital image data is input to the digital signal processor 230.

The digital signal processor 230 as the main controller processes the digital signal from the CDS-ADC element 210 to generate digital image data classified into luminance and chroma signals.

The video-signal generator 240 converts digital image data from the digital signal processor 230 into a video signal S _VID which is an analog image signal.

The digital signal processor 230 as the main control unit communicates with host devices, for example, computers, via a communication interface 260 and a communication channel D _COM , and video-signal via a video signal channel S _VID . The video signal from the generator 240 is transmitted to the host devices.

Meanwhile, the micro-computer 270 controls the driving unit 250 to control the aperture motor M _A , the zoom motor M _Z , the focus motor M _F , the filter motor M _D , and the panning motor M _P. And a tilting motor M _T. In addition, the micro-computer 270 controls the lighting unit 280 to illuminate the illumination light.

The aperture motor M _A drives the aperture, the zoom motor M _Z drives the zoom lens, and the focus motor M _F drives the focus lens. The filter motor M _D drives the optical low pass filter OLPF and the infrared cut filter IRF at the filter part.

The panning motor M _P rotates the optical system OPS from side to side. The tilting motor M _T rotates the optical system OPS up and down.

Hereinafter, recognition of the neural network skin color using normalized skin color data performed by the digital signal processor 230 as the main controller will be described with reference to FIGS. 3 and 4.

3 is a detailed block diagram of the digital signal processor 230 of FIG. 2.

Referring to FIG. 3, the digital signal processor 230 includes a buffer 231, a motion detection boom 232, a converter 233, a normalizer 234, a skin detector 235, a position calculator 236, and And a zoom control unit 237.

The motion detector 232 detects motion by comparing an input current (t) image with a previous (t-1) image stored in the buffer 231.

The converting unit 233 converts the RGB signal of the motion detected image into a YCbCr signal which is a luminance signal and a color difference signal.

The normalizer 234 normalizes the color signal CbCr of the detected motion image by using the luminance signal Y of the detected motion image, and the normalized output signal DbDr is represented by Equation 1 below.

In Equation 1, n is a first normalization factor, and when the Cr value is expressed as a negative number, n is a value shifted in order to express a positive number. M is a second normalization factor and is a coefficient value between 0-1 for generating an input value when using a backpropagation algorithm for skin detection described later. Since the color signal CbCr is divided by the luminance signal Y at the time of normalization, the influence of illuminance is removed at the time of skin detection using a backpropagation algorithm, so that the judgment boundary area is small, so that the skin recognition rate can be increased. The computational speed and learning time are reduced to improve the processing speed.

The skin detector 225 detects the skin by performing a backpropagation algorithm based on the weight obtained by inputting the normalization signal DrDb and prior learning.

4 is a diagram illustrating a weight generated through pre-learning with a back-propagation algorithm. Referring to FIG. 4, the backpropagation algorithm performing unit 235-1 is formed of a multilayer of an input layer, a hidden layer, and an output layer, and thus is useful when setting a nonlinear boundary region. In general, learning requires input data and output data. The input data is a value (DbDr (see FIG. 3)) which is previously normalized by converting the RBG signal of the input image into a YCbCr signal. When the input data DrDb is multiplied by the weight of the neural network (W _ji , W _kj ) and repeated, the result is outputted. _ji , W _kj ) is updated inversely to update the weight (W _ji , W _kj ). The output data proceeds in the direction of the input layer, the hidden layer, and the output layer, and the update of the weights W _ji and W _kj proceeds in the opposite directions of the output layer and the hidden layer. Since such a backpropagation algorithm is a well-known conventional technique, detailed description is omitted.

The skin detector 235 performs a backpropagation algorithm and outputs 1 when the skin color is detected in the corresponding video signal, and 0 otherwise.

The position calculator 236 counts the number of pixels detected by the skin and calculates the position.

If the number of pixels counted by the position calculator 236 is greater than or equal to a certain number, the zoom controller 237 outputs a zoom control signal to the driver 250 to perform zooming around the corresponding position. The driver 250 receives the zoom control signal to operate the zoom motor M _Z.

5 is a flowchart illustrating an operation of a digital image processing method according to another embodiment of the present invention.

The digital signal processor 230 detects a motion by comparing the current (t) image and the previous (t-1) image stored in the buffer 231 (step 501).

Subsequently, the digital signal processor 230 converts the RGB signal of the image from which the motion is detected into a YCbCr signal (step 503).

The digital signal processor 230 normalizes the color signal CbCr using the luminance signal Y, the first normalization factor, and the second normalization factor (step 505). The normalized signal DbDr is used as input data of the backpropagation algorithm.

Thereafter, the digital signal processor 23 detects the skin by performing a backpropagation algorithm based on the weight obtained by inputting the normalized signal DrDb and pre-learning (step 507). The digital signal processor 230 outputs 1 when the skin color is detected and 0 when it is detected.

When the skin color is detected, the digital signal processor 230 counts the number of pixels detected by the skin and calculates the position (step 509).

When the position is calculated, the digital signal processor 230 outputs a zoom control signal for zooming around the position to the driver 250, and the driver 250 receives the zoom control signal to receive the zoom motor M _Z. Start operation (step 511).

The specific acts described in the present invention are, by way of example, not intended to limit the scope of the invention in any way. For brevity of description, descriptions of conventional electronic configurations, control systems, software, and other functional aspects of such systems may be omitted. Also, the connections or connecting members of the lines between the components shown in the figures are illustrative of functional connections and / or physical or circuit connections, which may be replaced or additionally provided by a variety of functional connections, physical Connection, or circuit connections. Also, unless explicitly mentioned, such as " essential ", " importantly ", etc., it may not be a necessary component for application of the present invention.

The use of the terms " above " and similar indication words in the specification of the present invention (particularly in the claims) may refer to both singular and plural. In addition, in the present invention, when a range is described, it includes the invention to which the individual values belonging to the above range are applied (unless there is contradiction thereto), and each individual value constituting the above range is described in the detailed description of the invention The same. Finally, the steps may be performed in any suitable order, unless explicitly stated or contrary to the description of the steps constituting the method according to the invention. The present invention is not necessarily limited to the order of description of the above steps. The use of all examples or exemplary language (e.g., etc.) in this invention is for the purpose of describing the invention in detail and is not to be construed as a limitation on the scope of the invention, It is not. In addition, one of ordinary skill in the art appreciates that various modifications, combinations and changes can be made depending on design conditions and factors within the scope of the appended claims or equivalents thereof.

101,111,121: surveillance camera 2: host device
OPS: Optical System OEC: Photoelectric Converter
210: CDS-ADC 220: timing circuit
230: digital signal processor 240: video-signal generator
250: drive unit 260: communication interface
270: micro-computer 280: lighting unit
231: buffer 232: motion detection unit
233: conversion unit 234: normalization unit
235: skin detection unit 236: position calculation unit
237: zoom control

Claims (6)

a) detecting a movement of an input image;
(b) normalizing a color signal of the detected motion image by using the luminance signal of the detected motion image;
(c) detecting skin from the normalized signal using a backpropagation algorithm; And
and (d) zooming around the area detected by the skin. The method of claim 1, wherein in normalizing in the step (b),
And dividing a value obtained by adding a first normalization factor to the color signal by the luminance signal, and then multiplying a second normalization factor to output a normal signal. 3. The method of claim 2, wherein the first normalization factor is a bit shift coefficient that makes the color signal positive, and the second normalization factor is a value between 0-1 for generating the backpropagation algorithm input value. Image processing method. A motion detector for detecting a motion by using an input video signal and a previous video signal;
A normalizer for normalizing a color signal of the detected motion image by using the luminance signal of the detected motion image;
A skin detector for detecting skin from the normalized signal using a backpropagation algorithm; And
And a zoom controller configured to output a zoom performing signal centering on the area detected by the skin. The method of claim 4, wherein the normalization unit
And dividing a value obtained by adding a first normalization factor to the color signal by the luminance signal, and then multiplying a second normalization factor to output a normal signal. 6. The method of claim 5, wherein the first normalization factor is a bit shift coefficient that makes the color signal positive, and the second normalization factor is a value between 0-1 for generating the backpropagation algorithm input value. Image processing device.

Images