KR100916854B1 - System and method for generating thumnail images - Google Patents

Abstract

The present invention relates to a system for generating a thumbnail and a method for generating the same. More particularly, a system for generating a thumbnail image file by extracting a single static image that most characteristically reveals the contents of the file from a dynamic image file including a plurality of image frames. And a method of producing the same.

By using the thumbnail generation system and the method of the present invention, an image having a simple background and a simple letter is excluded from a dynamic image file including a plurality of image frames, and a high quality image which most reveals the contents of the file is characterized. The effect is that you can extract and create thumbnail image files.

Thumbnail, histogram, hue, saturation, color space, noise

Description

Thumbnail generation system and method of creation {SYSTEM AND METHOD FOR GENERATING THUMNAIL IMAGES}

The present invention relates to a system for generating a thumbnail and a method for generating the same. More particularly, a system for generating a thumbnail image file by extracting a single static image that most characteristically reveals the contents of the file from a dynamic image file including a plurality of image frames. And a method of producing the same.

Currently, most internet shopping mall sites provide their own product search functions. In addition, when a product is searched on an internet shopping mall site, the search result is generally output to include a list of search target products and a thumbnail image of each product.

Therefore, it is very important that the quality thumbnail image is exposed in the product search results of the Internet shopping mall site. This is because the neat thumbnail image can visually clearly convey the characteristics of the product, such as the shape or color of the product, and thus is easy to attract the user's attention. Among the Internet shopping mall sites, especially in the so-called online market place, which is an online market place where individual sellers sell their goods, such as auction or Gmarket, the customer attraction of the product image is high. This is particularly important in improving sales for individual sellers.

On the other hand, the thumbnail image of the Internet shopping mall may be uploaded directly in the form of a static image file (for example, a general image file such as JPEG, GIF) by the product seller or the Internet shopping mall manager. Alternatively, a specific still image image may be automatically extracted and registered on an Internet shopping mall site from a dynamic image file uploaded by a product seller, for example, a file in a dynamic GIF format.

That is, the merchandise seller uploads a dynamic image file including a plurality of images of the merchandise in order to attract the gaze of the Internet shopping mall user. Such a dynamic image file is displayed alternately in a predetermined order a plurality of images contained in the dynamic image file in a specific area in the web page about the product. These dynamic image files typically include images that show the product's appearance, and many of the dynamic image files used on Internet shopping mall sites contain text such as "Sale," "Free Shipping," and images about the product's appearance. Etc. are included.

On the other hand, many Internet shopping mall sites extract thumbnails from the dynamic image files uploaded by product sellers to generate thumbnail images and display them together on the product search result list screen. There are many reasons for displaying thumbnail images created by extracting still images, rather than displaying dynamic image files uploaded by commodity sellers, but among them, the main purpose is to reduce traffic between internet shopping mall sites and users.

One method used to extract a specific still image file from the dynamic image file is to extract the first frame of the plurality of image frames included in the dynamic image file.

For example, a method of extracting only the first image frame among the plurality of image frames included in the dynamic GIF image and then resizing it to the size of the desired thumbnail image is used.

In this method, since the image of the first frame of the dynamic GIF file is unconditionally selected, a separate calculation process is unnecessary and thus the speed is high. On the other hand, regardless of the content of the image, the image of the first frame is uniformly selected, so if the image of the first frame of the dynamic GIF file does not adequately represent the meaning of the product or is meaningless, the thumbnail image file also appropriately expresses the characteristics of the product. There is a problem of not being able or meaningless.

Accordingly, a system for excluding a image having a simple background and a simple letter from a dynamic image file including a plurality of image frames, and extracting a high quality image that most characteristically shows the contents of the file to generate a thumbnail image file And a request for a generating method have been steadily raised.

The present invention was conceived in response to the above-described request, and excludes an image having a simple background and a simple letter from a dynamic image file including a plurality of image frames, and provides a high quality image that most characteristically reveals the contents of the file. It is an object of the present invention to provide a system and a method for generating the same by extracting a thumbnail image file.

In order to achieve the above object, a thumbnail generation system according to the present invention includes a color space conversion unit for converting color spaces of a plurality of image frames included in a dynamic image file, and calculating a color histogram for each of the image frames. A color histogram analyzing unit, a frame selecting unit which selects one of the plurality of image frames as an optimal image frame based on the color histogram for each of the image frames, and a thumbnail that converts the optimal image frame into a thumbnail image file It includes a generation unit.

In addition, the thumbnail generation method according to the present invention includes a color space conversion step of converting the color space of the plurality of image frames included in the dynamic image file,

A color histogram analysis step of calculating a color histogram for each of the image frames; a frame selection step of selecting one of the plurality of image frames as an optimal image frame based on the color histogram for each of the image frames; And a thumbnail generation step in which the optimum image frame is converted into a thumbnail image file.

Using the thumbnail generation system and the method of the present invention, it is possible to generate a thumbnail image file by extracting a single static image that most characteristically reveals the contents of the file from a dynamic image file including a plurality of image frames. have.

DETAILED DESCRIPTION Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art may easily implement the present invention. As those skilled in the art would realize, the described embodiments may be modified in various different ways, all without departing from the spirit or scope of the present invention. In the drawings, parts irrelevant to the description are omitted in order to clearly describe the present invention, and like reference numerals designate like parts throughout the specification.

Throughout the specification, when a part is "connected" to another part, this includes not only "directly connected" but also "electrically connected" with another element in between. .

Throughout the specification, when a part is said to "include" a certain component, it means that it can further include other components, without excluding other components unless specifically stated otherwise. In addition, the terms “… unit”, “… unit”, “module”, etc. described in the specification mean a unit that processes at least one function or operation, which may be implemented by hardware or software or a combination of hardware and software. have.

1 is a block diagram showing an example of a thumbnail generation system according to the present invention.

As shown in FIG. 1, the thumbnail generation system 10 according to the present invention includes an image reader 100, a noise filter unit 110, a color space converter 120, a color histogram analyzer 130, and frame selection. The unit 150 and the thumbnail generation unit 160 is made.

The summary of the operation of the thumbnail generation system 10 according to the present invention is as follows.

First, the thumbnail generation system 10 receives a dynamic image file (eg, a dynamic GIF file) having a plurality of image frames.

The image reading unit 100 classifies and reads image data included in the dynamic image file for each image frame.

Next, the noise filter 110 performs noise reduction filtering on the image data of each image frame.

Next, the color space converter 120 converts the color space of each image data from the original color space (eg, RGB) to another color space (eg, HSV).

Next, the color histogram analyzer 130 generates a color histogram of the image data for each image frame, and analyzes the result value of the generated color histogram.

Next, the frame selector 150 selects an image frame most suitable for generating a thumbnail image by combining the analysis results performed by the color histogram analyzer 130 and the chroma histogram analyzer 140.

Finally, the thumbnail generator 160 generates the selected image frame as a static image file having the required resolution and size.

A detailed description of the roles and operations of the noise filter 110, the color space converter 120, and the color histogram analyzer 130 among the components of the thumbnail generation system 10 mentioned above is as follows. .

First, the noise filter 110 plays a role of reducing noise in order to reduce measurement errors when measuring the degree of color distribution of an image included in each image frame.

The thumbnail generation system 10 according to the present invention measures and compares how much color is included in each image frame. In order for the comparison to be meaningful, it is necessary to accurately grasp the type and distribution of colors included in the image frame.

Therefore, if there is a lot of noise in an image, it will become difficult to obtain an accurate distribution. In this case, the noise included in the image means color information, not color information necessary to express the corresponding image. That is, there is a relationship in which an image containing a lot of noise is recognized as a "dirty" image even with the naked eye.

Such noise is often unintentionally included in the process of producing a dynamic image file, and it is necessary to remove the noise to obtain a higher quality image.

Therefore, the noise filter 110 performs noise filtering on the image included in each image frame.

On the other hand, there are a variety of filtering methods for reducing noise in an image. In the present embodiment, a smoothing process is performed using a commonly used Gaussian filter.

However, the noise filter of the noise filter unit 100 of the present invention is not limited to a Gaussian filter, and any filter that can be used appropriately to reduce noise included in an image may be widely applied according to any algorithm.

Next, the color space converter 110 converts a color space of each image into another color space.

A digital image consists of a large number of pixels, and each pixel contains information about each of the three primary colors of light, red (R), green (G), and blue (B). Therefore, the digital image may be regarded as a collection of data having information values indicating color.

In addition, the color space refers to a color reproduction source necessary for expressing a color based on the information values representing the above colors.

As for the color space, there are various standards such as sRGB, AdobeRGB, and HSV. In particular, the sRGB color space adopted by the Microsoft Windows operating system which is widely used around the world is common.

By the way, the images in the RGB color space we are actually seeing (including both sRGB and AdobeRGB, are the same below) are significantly affected by lighting.

That is, even though the color of the subject is the same, the R value, the G value, and the B value all change according to the illuminance of the illumination shining on the subject. For example, the R value of an image taken in an environment with low illuminance of the same subject does not match the R value of an image taken in an environment with high illuminance, and the G value and B value are also the same. Due to this mismatch, the color histogram characteristics and the color histogram characteristics of the two images are also different.

On the other hand, in the HSV color space, color (H: hue) and saturation (S: saturation) and brightness factors (V: value) are separated, thereby reducing the influence of light and shade on the image. There is an advantage.

Here's how to convert an image from an RGB color space to an HSV color space:

Convert from RGB color space to HSL or HSV color space

(r, g, b) is called red, green and blue coordinates for a specific color. At this time, each of the red, green, and blue coordinate values is a real number between 0 and 1. In addition, max is the same value as the maximum value of r, g, b, and min is the same value as the minimum value of r, g, b.

h ∈ [0, 360) is the hue angle in degrees, s ∈ [0,1] is the saturation, and l ∈ [0,1] is the brightness, in the HSL color space ( h , s , l ) values are defined as in Equation 1 below.

The value of h is normally normalized between 0 degrees and 360 degrees, with h = 0 being assumed when max = min (eg, gray).

In addition, the definition of the color in the HSL color space and the HSV color space is the same, but the saturation (s) and brightness (v) is defined in Equation 2 below in the HSV color space.

Color histogram analyzer,

The color histogram analyzer 130 calculates a color histogram of each image frame in order to filter an image having a simple letter on a simple background.

2A is an example of an image having monochrome letters on a monochrome background.

As shown in FIG. 2A, an image having monochrome letters on a monochrome background is very unlikely to be an image representing the shape of the product itself. Therefore, as shown in FIG. 2, an image having a monochromatic letter on a solid background should be excluded from being selected as a thumbnail image.

Other examples of images to be excluded from being selected as thumbnail images include images having text of various colors, images having text and complex backgrounds, and images having only a single color background.

In the present invention, the color histogram analyzer 130 filters such an image.

FIG. 3A is an example of an image frame having color photograph images of various colors, FIG. 2B is a view showing the color histogram of FIG. 2A, and FIG. 3B is a view showing the color histogram of FIG. 3A.

If the specific dynamic image file includes FIGS. 2A and 3A as image frames, the color histogram analyzer 130 may use the color histogram as shown in FIG. 2B for FIG. 2A and the color as shown in FIG. 3B for FIG. 3A. Generate each histogram.

In the color histograms shown in Figs. 2B and 3B, the horizontal axis represents the color represented by the pixel, and the vertical axis represents the frequency of the pixel having each color in the corresponding image frame. Due to the characteristics of the HSV color space, the horizontal axis has a value from 0 degree to 180 degrees.

The image shown in FIG. 2A is an image in which monochrome characters (white) appear on a monochrome background (red). Referring to FIG. 2B, which is the color histogram of FIG. 2A, the pixel frequency of the red color gamut 20 is remarkably high, and the colors 21 and 22 adjacent to the red are partially observed.

At this time, the number of color columns that intersect the horizontal axis 24 (hereinafter referred to as "reference horizontal axis") corresponding to a predetermined pixel frequency (hereinafter referred to as "reference pixel frequency f") having a sufficiently high frequency of pixels is It becomes one about the red column 20.

In other words, it means "one color having a pixel frequency higher than the reference pixel frequency f".

On the other hand, the image shown in FIG. 3A is assumed to be a color photograph image of various colors (although it is represented by a halftone image). Referring to FIG. 3B, which is the color histogram of FIG. 3A, it can be seen that a color having a high frequency of pixels is observed in various color regions when compared to FIG. 2B.

In this case, as in FIG. 2B, the number of columns 30, 31, 32, 33, and 34 that intersect the reference horizontal axis 35 is five.

In this manner, the color histogram analyzer 130 calculates the number of columns that cross the reference horizontal axis for each image frame of the input dynamic image file.

On the other hand, according to the results obtained by repeated experiments, in the case of an image frame having a simple background and simple letters, the number of color columns that cross the reference horizontal axis is about 4 to 5 or so.

Therefore, in the case of an image frame having a simple background and simple letters, the number of pixel frequencies corresponding to the reference horizontal axis is about 4 to 5 color columns that cross the reference horizontal axis. It can be set to be a pixel frequency within a range such that the number of intersecting color columns is significantly greater than this.

Frame selector,

The frame selector 150 selects an optimal thumbnail image through the number of color columns of each image frame calculated by the color histogram analyzer 130.

First, a voting is performed by comparing the number of color columns of each image frame calculated by the color histogram analyzer and assigning a weight to each image frame.

In this case, the highest weight is assigned to the image frame having the largest number of color columns, and the lower weight is assigned to the image frame having the smaller number of color columns.

As described above with respect to the color histogram analyzer, it may be determined that an image frame given the highest weight by bowing has a high probability of being an image frame including the most colorful image. Similarly, it may be determined that the lowest weighted image frame has a high probability of being a simple color background or a simple color background.

Accordingly, the frame selector 150 finally selects the highest weighted image frame as the optimal image frame to generate a thumbnail image through bowing.

Thumbnail generation unit

When the optimal image frame is selected by the frame selector 150, the thumbnail generator converts the image frame into a thumbnail image having a desired size, resolution, and file format to generate a thumbnail image file as a final result.

The method of generating the image file is not limited to a specific method, and any method currently existing in connection with the conversion of the image file may be widely applied.

4 is a flowchart illustrating an example of a thumbnail generation method according to the present invention.

As shown in FIG. 4, the thumbnail generation method according to the present invention includes an image reading step S100, a noise filtering step S110, a color space conversion step S120, a color histogram analysis step S130, Frame selection step S150 and thumbnail generation step S160.

First, in the image reading step (S100), the thumbnail generation system receives a dynamic image file having a plurality of image frames, and then reads the image data included in the dynamic image file separately for each image frame.

Next, in the noise filtering step S110, noise filtering is performed on each image frame. The noise filtering may be performed by a smoothing process using a Gaussian filter, but the present invention is not limited to noise filtering by this method. Any filter that can be used to reduce noise included in an image may be variously applied. .

In the color space conversion step (S120), the RGB color space of the input dynamic image file is converted into another color space.

For example, it may be converted from the RGB color space to the HSV color space according to the conversion method defined in Equations 1 and 2.

On the other hand, the color space conversion step is not limited to the conversion to the HSV color space, since the purpose is to convert the color space so that the change in contrast can have a relatively small effect on the image compared to the RGB color space. It may be possible to convert to a matching color space, such as the HSL color space.

In the color histogram analysis step S130, a color histogram for each image frame of the dynamic image file is calculated.

At this time, the number of color columns that intersect the reference horizontal axis for each image frame is calculated together.

In other words, in the color histogram analysis step S130, the number of colors larger than the reference pixel frequency is calculated.

Through the color histogram analysis step S130, an image frame having a small number of color columns that intersect the reference horizontal axis, that is, an image frame having a simple background and a simple letter may be found and filtered.

In the frame selection step S150, an optimal image frame to be generated as a thumbnail image is selected through the number of color columns of each image frame calculated in the color histogram analysis step S130.

The frame selection step S150 may further include a bowing step S152.

In the bowing step S152, weights are assigned to each image frame based on the number of color columns of each image frame calculated by the color histogram analyzer.

In this case, the highest weight is assigned to the image frame having the largest number of color columns, and the lower weight is assigned to the image frame having the smaller number of color columns.

The highest weighted image frame in the bowing step is selected as an optimal image frame to be generated as a thumbnail image.

In the thumbnail generation step S160, the optimal image frame selected in the frame selection step S150 is converted into a thumbnail image having a desired size, resolution, and file format.

The embodiments of the present invention described above are not implemented only through the apparatus and the method, but may be implemented through a program for realizing a function corresponding to the configuration of the embodiment of the present invention or a recording medium on which the program is recorded. Implementation may be easily implemented by those skilled in the art from the description of the above-described embodiments.

Although the embodiments of the present invention have been described in detail above, the scope of the present invention is not limited thereto, and various modifications and improvements of those skilled in the art using the basic concepts of the present invention defined in the following claims are also provided. It belongs to the scope of rights.

1 is a block diagram showing an example of a thumbnail generation system according to the present invention;

2A is an example of an image having monochrome letters on a monochrome background,

2b is a color histogram of FIG. 2a;

3A is an example of an image frame having color photographic images of various colors;

3b is a color histogram of FIG. 3a;

4 is a flowchart illustrating an example of a thumbnail generation method according to the present invention.

Claims (11)

A color space conversion unit for converting color spaces of a plurality of image frames included in the dynamic image file so that color and brightness elements are separated; A color histogram analyzer for calculating a number of colors having a pixel frequency greater than a color histogram and a reference pixel frequency for each of the image frames; A frame selection unit for selecting one of the plurality of image frames as an optimal image frame based on the color histogram for each of the image frames; A thumbnail generator for converting the optimal image frame into a thumbnail image file, The frame selector assigns a weight to each of the image frames based on the number of colors, wherein the weights given to the image frames with a greater number of colors are higher than the weights given to the image frames with fewer colors. Thumbnail generation system. The method of claim 1, The thumbnail generation system further comprises a noise filter for performing noise reduction filtering on each of the image frames. delete The method of claim 1, And the frame selection unit selects an image frame having the highest weight among the plurality of image frames as the optimal image frame. A color space conversion step of converting color spaces of a plurality of image frames included in the dynamic image file to separate color and brightness elements; A color histogram analysis step of calculating a number of colors having a pixel frequency greater than a color histogram and a reference pixel frequency for each of the image frames; A bowing step in which each of the image frames is weighted based on the number of colors; A frame selection step of selecting one of the plurality of image frames as an optimal image frame based on the color histogram for each of the image frames; and A thumbnail generation step of converting the optimal image frame into a thumbnail image file, In this case, the thumbnail generation method so that the weight given to the image frame having a large number of colors is higher than the weight given to the image frame having a smaller number of colors. The method of claim 5, And a noise filtering step in which noise attenuation filtering is performed on each of the image frames, before the color histogram analysis step. delete delete delete delete The method of claim 5, And in the frame selection step, a thumbnail image frame having the highest weight among the plurality of image frames is selected as the optimal image frame.

Images