JPH07129762A - Sketch-fashion image generator - Google Patents

Abstract

PURPOSE:To generate a binary picture provided with feature of sketch fashion by sampling a main contour line and shading information from a natural picture, and generating a sketch-fashion picture from sampled contour line and shading information. CONSTITUTION:A contour line sampling means 4 performs the quantitative determination of local gradients in four adjacent directions from an inputted multilevel original picture 217, and generates a contour line picture 213 set as a value which means the contour line when a quantitative determination value exceeds a first threshold value, and a value which means a background in the case other than that. A shading sampling means 102 generates a shading picture 215 depending on shading expression set as a value in which a part where the pixel value of the original picture is smaller than a second threshold value means shade, and a value which means the background in the case other than that. A picture synthesizing means 103 inputs the contour line picture 213 and the shading picture 215, and outputs by generating the sketch-fashion picture 218 set as a value which means the value of the background when both picture element values of those pictures at the same coordinate position means the background, and a value which means a diagram in the case other than that.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus,
In particular, it relates to a binarization method of multi-valued images and an apparatus thereof.

[0002]

2. Description of the Related Art As a method of binarizing a natural image such as a landscape photograph, there are an error diffusion method and a shading method [edited by Mikio Takagi / Yoshihisa Shimoda, "Image Analysis Handbook", pp.
493-pp. 501, University of Tokyo Press (199
1)]. These methods are intended for pseudo halftone expression, and their main purpose is faithful expression of natural images.

However, for a natural image mainly composed of straight lines such as mechanical parts, it is preferable to represent it as a sketch style composed of lines and shadows, rather than generating a binary image by the above-described binarization method. However, it is more practical and practical.

As a technique for generating a sketch-like image, the same object is imaged under a lighting condition that does not produce a shadow due to its shape and an illumination condition that causes a shadow due to that shape, and three-dimensional is obtained by calculating each image data obtained. A method has been proposed in which information is acquired and an illustration image is created based on the information [JP-A-1-281574].

[0005]

In the above-described illustration creating method, it is a principle that an image is picked up under limited illumination conditions, and for an image already taken under arbitrary illumination conditions, a preferable sketch-like image is obtained. There is no guarantee that you will get it.

Further, the above-mentioned illustration creating method aims at three-dimensionally expressing the shape of an object by an illustration, and removes the color pattern component and the texture component on the surface of the object. Therefore, an image such as a portrait cannot express the pattern and texture of clothes.

The method and apparatus of the present invention can generate a binary image having sketchy features from an arbitrary image, and can generate and modify a binary image having sketchy features in a short time. The purpose is to provide an interactive environment.

[0008]

In order to achieve the above object, a sketch-like image generating apparatus according to a first aspect of the present invention includes means for extracting contour information from an original image having multiple gradations, and means for extracting contour information from the original image. A means for extracting shadow information, and a synthesizing means for reconstructing an image from the contour information and the shadow information,
It is characterized in that an image having features of rough drawing is generated.

The sketch-like image generating apparatus according to the second invention is
In order to achieve the above object, a first storage unit that stores an original image having multiple gradations, and at least one of a first threshold value and a second threshold value is acquired from an operator and output, or registered. Threshold value input means for outputting at least one of the first threshold value and the second threshold value, and the first threshold value as an input, quantifying the gradient of the local pixel value of the original image,
A contour line extracting means for generating and outputting a contour line image composed of a contour line and a background using the quantitative value and the first threshold value;
A shadow extraction unit that receives the second threshold as an input, generates a shadow image including a shadow portion and a background by using the original image, the second threshold, and a shadow expression method, and stores a second contour image. Storage means, a third storage means for storing the shadow image, the contour line image stored in the second storage means, and a sketch style from the shadow image stored in the third storage means. An image synthesizing means for generating an image, a fourth storage means for storing the sketch-like image, an image display means for displaying the sketch-like image in the fourth storage means, and a correction for the sketch-like image. An environment for designating a policy is provided, and modification input means for outputting the modification information, and the modification information as an input, and the sketch-like image in the fourth storage means is modified according to the modification information to modify the sketch-like image. Image correction hand to write back to the storage means When the has, and generating an image having the characteristics of Drawings image from the original image.

According to a third aspect of the present invention, there is provided a sketch-like image generating apparatus.
In order to achieve the above object, first storage means for storing an original image having multiple gradations, second storage means for storing a sketch-like image in which pixels are initialized as an initial state, and first storage means. Threshold input means for acquiring and outputting at least one of a threshold and a second threshold from an operator, or outputting at least one of the registered first threshold and second threshold,
Using the first threshold as an input, the gradient of the local pixel value of the original image is quantified, and the outline is stored in the second storage means using the quantitative value and the first threshold. The outline extraction unit that is generated in the image and the second threshold value are input, and the original image, the second threshold value, and the shadow expression method are used to draw the shaded portion in the fourth storage unit. A shadow extraction unit that occurs in the first storage unit, an image display unit that displays the sketch-like image in the second storage unit, and an environment that specifies a policy for correcting the sketch-like image, and outputs the correction information. Correction input means; and image correction means for inputting the correction information and correcting the sketch-like image in the second storage means according to the correction information and writing back the image in the second storage means. , Generate an image with features of raw drawing from the original image Characterized in that it.

[0011]

A feature of sketch drawing and drawing drawing is that a drawing object is expressed by a contour line and a shadow expression method. Therefore, converting a natural image such as a landscape photograph or a portrait photograph so as to give an observer the impression of a raw drawing requires the following procedure. Step 1: Extract main contour lines from the natural image. Step 2: Extract main shadow information from the natural image. Step 3: A sketch-like image is generated from the extracted contour line and shadow information. However, the procedure 1 and the procedure 2 are processes that can be performed independently, and therefore the order is not specified.

An image signal group represented by a two-dimensional array obtained by an image input means such as a scanner will be referred to as an original image. Further, one image signal is called a pixel, and the value of the image signal is called a pixel value. Regarding the pixel value of the original image, 0 means black, and as the pixel value increases, the brightness increases to dark gray, gray, and light gray, and when the pixel value reaches the maximum, it means white.

The procedure 1 corresponds to a process of generating an edge image from an original image and obtaining a main contour line from the edge image.

A method of generating an edge image will be described with reference to FIG. When an arbitrary coordinate of the original image is set as a pixel of interest 40, among the eight pixels 41 to 48 adjacent to this pixel, an upper left pixel 41 and a lower right pixel 45, an upper right pixel 43 and a lower left pixel 47, and a right adjacent pixel 44, the pixel 48 on the left, and the pixel 4 above
The absolute value of the difference between the pixel values of four pairs of diagonal pixels of 2 and the lower pixel 46 is calculated, and the largest value is set as the pixel value of the target pixel 40. Alternatively, as shown in Expression 1, the sum of the absolute values of the differences between the pixel values of the four pairs of diagonal pixels may be used as the pixel value of the target pixel 40.

D (x, y) = | S (x-1, y-1) -S (x + 1, y + 1) | + | S (x + 1, y-1) -S (x-1, y + 1) | + | S (x, y-1) -S (x, y + 1) | + | S (x-1, y) -S (x + 1, y) | (Equation 2) If this calculation is performed for all pixels of the original image, , An edge image having all positive edge components is obtained.

If the magnitude of the gradient of the pixel value around the target pixel can be quantified and the calculation is performed so that the value becomes an absolute value, the same effect as that of the present invention can be obtained. An image generation method may be used instead.

A portion having a large pixel value in the edge image means a boundary at which there is an abrupt change in luminance in the original image, and thus it may be considered as a main contour line. Therefore, by setting an appropriate threshold value, the pixel value smaller than the threshold value is 1
In addition, the edge image is binarized by a threshold process in which the pixel value equal to or more than the threshold value is set to 0, and the edge image is obtained. Therefore, in the contour image, the pixel value 1 means the background and the pixel value 0 means the contour line.

The procedure 2 corresponds to extracting a dark portion (that is, a portion having a small pixel value). Therefore,
An appropriate threshold value is provided, and if the pixel value of the original image is smaller than the threshold value, it is set to 0, and otherwise, it is set to 1 and binarized to obtain a shadow image.

Further, the shadow portion may be expressed by a method other than the above solid-painting expression. For example, in a region where the pixel value of the original image is smaller than the threshold value, the shadow can be expressed by a stippled expression or a hatched expression according to the gradation value. The point is that a pixel value 1 means a background, and a pixel value 0 means a shadow image which means a shadow portion.

The threshold value is set so that the area of the shadow portion (that is, the portion having a pixel value of 0) is about 10% of the entire area. The threshold value that satisfies this condition can be easily obtained by taking the histogram of the pixel values of the original image. That is, when the frequency of the histogram is accumulated from the smaller pixel value, the pixel value at which the accumulated result becomes 10% of the total number of pixels is adopted as the threshold value.

The procedure 3 corresponds to performing image composition of the contour line image and the shadow image. When the contour image and the shadow image are superposed, the logical product of the pixel values at the same coordinate positions is taken as the pixel value of the composite image. Since the operation of the logical product is 1 and 1 = 1 1 1 and 0 = 0 0 and 1 = 0 0 and 0 = 0, the image composition in which the contour line in the contour line image and the shadow part in the shadow image are respectively saved You can
It is important that the composite image preserves the contour lines and shaded areas. With this property, there is an advantage that in the image correction processing described later, the contour line and the shaded portion can be simultaneously and intuitively corrected without performing them individually.

When the composite image is displayed on the display,
By assigning white to the portion having the pixel value 1 and assigning black to the portion having the pixel value 0, it is possible to display a sketch-like image as if drawn on a white paper with magic. When it is desired to output a hard copy to a printer or the like, it suffices to attach ink only to the portion having the pixel value of 0.

By the above procedure, the natural image can be converted into a sketch-like image. However, by performing the correction processing of the contour line and the shadow on the composite image, a more preferable sketch-like image can be obtained. it can.

As the above-mentioned correction process, a synthetic image is displayed on a display, and the operator uses an input device such as a pointing device to add or delete a figure such as a point, a line, or a surface, thereby interactively correcting the figure. Method, there is an automatic method of removing noise and connecting broken lines by utilizing image processing technology. As an example of the latter, the simplest method is the expansion / contraction process of a figure [edited by Mikio Takagi / Yoshihisa Shimoda, "Image Analysis Handbook", pp. 57
5-pp. 576, The University of Tokyo Press (1991)].

[0025]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An illustration image generating apparatus according to a first embodiment of the present invention will be described with reference to FIG. This apparatus receives an original image 217 having multiple gradations, converts it into a rough drawing-like image 218 having characteristics of rough drawing, and outputs it.

The contour line extraction means 101 quantifies local gradients in the four adjacent vertical, horizontal, and diagonal directions from the input original image 217, and when the quantitative value is larger than the first threshold value, the contour is extracted. A contour line image 213 having a value indicating a line and a value indicating a background in other cases is generated and output.

The shadow extracting means 102 uses the original image 217.
A shadow image 215 is generated and output depending on a shadow expression method in which a portion having a pixel value of is smaller than the second threshold value means a shadow, and in other cases, a value meaning a background.

The image synthesizing means 103 receives the contour line image 213 which is the output of the contour line extracting means 101 and the shadow image 215 which is the output of the shadow extracting means 102, and which receives the contour line image 213 and the shadow image. If the pixel values at the same coordinate positions of 215 both indicate the background, the background value is generated, and in other cases, the sketch-like image 218 is generated and output.

A drawing style image generating apparatus according to a second embodiment of the present invention will be described with reference to FIG. This apparatus converts the multivalued original image stored in the original image storage means 200 into a sketch-like image having the characteristics of the sketch-like drawing, and draws the sketch-like image storage means 21.
It is stored in 0. Further, by modifying the sketch-like image in the sketch-like image storage means 210, it is changed to a more preferable sketch-like image, and the sketch-like image storage means 21.
It is stored in 0.

The contour line threshold value input means 201 interactively acquires the contour line threshold value 212 from the operator and outputs the contour line threshold value 212 as a criterion for binarizing the contour line. If not specified by the operator, the contour threshold 212 registered in advance is output.

The contour line extraction means 202 receives the contour line threshold value 212 output from the contour line threshold value input means 201 as an input, and the original image 2 stored in the original image storage means 200.
The local gradients in the four adjacent vertical, horizontal, and diagonal directions of 17 are quantified. If this quantified value is larger than the contour line threshold value 212, it means a contour line, and in other cases, it means a background. The contour image 213 having the value is generated and output.

The contour line image storage means 203 stores the contour line image 213 output from the contour line extraction means 202.

The shadow threshold inputting means 204 interactively acquires and outputs a shadow threshold 214, which is a criterion for the shadow portion, from the operator. If not specified by the operator, the shadow threshold value 214 registered in advance is output.

The shading extraction means 205 receives the shading threshold value 214, which is the output of the shading threshold value input means 204, and receives the original image 217 in the original image storage means 200.
A shadow image 215 is generated and output depending on a shadow expression method in which a pixel value of is smaller than the shadow threshold value 214 represents a shadow value, and otherwise a background value.

The shadow image storage means 206 stores the shadow image 215 which is the output of the shadow extraction means 205.

The image synthesizing means 207 stores the contour lines in the contour line storage means 203 at the time when new image data is stored in either or both of the contour line image storage means 203 and the shadow image storage means 206. When the pixel values of the image 213 and the shadow image 215 in the shadow image storage unit 206 at the same coordinate position both indicate the background, the background value is set, and otherwise the line drawing is set. The image 218 is generated and output.

The sketch-like image storage unit 210 stores the sketch-like image 218 output from the image combining unit 207.

The image display means 211 presents the sketch-like image generation result to the operator by displaying the sketch-like image 218 stored in the sketch-like image storage means 210.

The correction input means 208 gives an instruction to add and delete graphic elements such as points, lines and planes to the sketch-like image 218 displayed on the image display means 211, or an automatic image correction process. Is provided and the correction information 216 is output.

The image correction means 209 receives the correction information 216, which is the output of the correction input means 208, as input, and according to the correction information 216, the sketch-like image storage means 21.
The sketch-like image 218 in 0 is corrected and stored in the sketch-like image storage means 210.

A sketch-like image generating apparatus according to a third embodiment of the present invention will be described with reference to FIG. In this apparatus, a sketch-like image initialized with a value indicating a background is stored in advance in the sketch-like image storage means 210, and the original image storage means 200 is stored.
Outlines and shaded parts extracted from the multivalued original image in
The sketch-like image having the characteristics of the sketch-like image is generated by adding to the sketch-like image of the sketch-like image storage means 210 as needed. Further, by modifying the sketch-like image of the sketch-like image storage means 210, a more preferable sketch-like image is changed, and the sketch-like image storage means 210 is changed.
Is something to remember.

Contour line threshold input means 201, shadow threshold input means 204, correction input means 208 and image correction means 209
The image display means and the image display means are the same as those in the second embodiment, and the description thereof is omitted.

The sketch-like image storage means 210 stores a sketch-like image in which all pixels are initialized with values indicating the background as an initial state.

The contour line extraction means 302 receives the contour line threshold value 212 output from the contour line threshold value input means 201 as an input, and the original image 217 in the original image storage means 200.
The local gradients in the four adjacent directions of the vertical, horizontal and diagonal directions are quantified, and a portion whose quantitative value is larger than the contour line threshold value 212 is set as a contour line, and the contour line is drawn by the sketch-like image storage means 210. Write on the wind image.

The shadow extraction means 305 receives the shadow threshold value 314 which is the output of the shadow threshold value input means 304 as an input, and outputs all the pixels of the original image 217 in the original image storage means 200 to the original image 217. When the pixel value is smaller than the shadow threshold value 214, the portion is defined as a shadow portion, and the shadow portion is subjected to a shadow expression technique to draw the sketch-like image storage means 2
Write in 10 sketch style images.

[0046]

As described above, the apparatus of the present invention generates an image having a feature of a drawing style image expressing a drawing object by a contour line and a shading expression method from an original image by a simple method or apparatus. It is especially effective when the information is easy to convey when expressed in a sketch style.

In the second embodiment, the contour line image 213
Since the shadow image 215 and the shadow image 215 are stored separately, the contour image 2
13 or one of the shaded images 215 is updated,
Immediately, the sketch-like image 218 can be synthesized and further displayed on the display means 211. Therefore, when the operator determines the contour line threshold value 212 using the contour line threshold value input unit 201, the display unit 211 always displays the sketch-like image in which the shadow is also added.
You can work while checking. Similarly, when the operator determines the shadow threshold 214 using the shadow threshold input means 204, the display means 2 always displays a sketch-like image with a contour line added.
You can work while checking in 11.

In the third embodiment, as an initial state, a sketch-like image meaning a blank sheet is prepared in the sketch-like image storage means 210, and the contour line extracting means 302 stores only the contour lines.
Since the shadow extracting means 305 is a method of writing only the shadow information into the sketch-like image, the contour line image storing means 203, the shadow image storing means 206, and the image synthesizing means 207 which are required in the second embodiment are unnecessary, and the device Is simplified.

[Brief description of drawings]

FIG. 1 is a block diagram of a first embodiment of the present invention.

FIG. 2 is a block diagram of a second embodiment of the present invention.

FIG. 3 is a block diagram of a third embodiment of the present invention.

FIG. 4 shows a positional relationship of pixels to be a target of calculation for obtaining a gradient of pixel values.

[Explanation of symbols]

 101 Contour Extraction Means of First Embodiment 102 Shade Extraction Means of First Embodiment 103 Image Compositing Means of First Embodiment 200 Original Image Storage Means 201 Contour Line Threshold Input Means 202 Contours of Second Embodiment Line extraction means 203 Contour image storage means 204 Shade threshold value input means 205 Shade extraction means of the second embodiment 206 Shade image storage means 207 Image synthesis means 208 of the second embodiment Correction input means 209 Image correction means 210 Drawing style Image storage means 211 Image display means 212 Contour line threshold 213 Contour line image 214 Shade threshold value 215 Shade image 216 Correction information 217 Original image 218 Sketch-like image 302 Contour line extraction means 305 of the third embodiment Shade of the third embodiment Extraction means 40 Target pixel 41 Pixel at upper left of target pixel 42 Pixel above target pixel 43 Upper right of target pixel Pixel 44 the pixel next to the right of the pixel of interest 45 the pixel lower right of the pixel of interest 46 the pixel below the pixel of interest 47 the pixel lower left of the pixel of interest 48 the pixel next left of the pixel of interest

Continuation of front page (51) Int.Cl. ⁶ Identification number Office reference number FI Technical indication location H04N 1/403 8420-5L G06F 15/66 A 4226-5C H04N 1/40 103 A

Claims (3)

[Claims] 1. A means for extracting contour information from an original image having multiple gradations, a means for extracting shadow information from the original image, and a synthesizing means for reconstructing an image from the contour information and the shadow information. A sketch-like image generating apparatus characterized by generating an image having a characteristic of rough drawing. 2. A first storage means for storing an original image having multiple gradations, and at least one of a first threshold value and a second threshold value acquired from an operator and output, or registered. Threshold input means for outputting at least one of a first threshold value and the second threshold value; quantifying a gradient of a local pixel value of the original image with the first threshold value as an input, and the quantitative value and the A contour line extracting unit that generates and outputs a contour line image composed of a contour line and a background using a first threshold value, and the original image, the second threshold value, and a shadow expression method using the second threshold value as an input. Shadow extraction means for generating a shadow image composed of a shadow portion and a background using, second storage means for storing the contour line image, third storage means for storing the shadow image, and second The contour line image stored in the storage means, and the third storage hand An image synthesizing unit that generates a sketch-like image from the shaded images stored in a row, a fourth storing unit that stores the sketch-like image, and an image that displays the sketch-like image in the fourth storage unit. Providing an environment for specifying a display means and a policy for modifying the sketch-like image,
Correction inputting means for outputting the correction information; and the fourth drawing means for correcting the sketch-like image in the fourth storing means by inputting the correction information.
An image correction device for writing back to the storage device, and generating an image having a characteristic of rough drawing from the original image. 3. A first storage means for storing an original image having multiple gradations, a second storage means for storing a sketch-like image in which pixels are initialized as an initial state, a first threshold value and a first threshold value. Threshold value input means for acquiring and outputting at least one of two threshold values from an operator or outputting at least one of the registered first threshold value and the registered second threshold value; and inputting the first threshold value Then, the gradient of the local pixel value of the original image is quantified, and the contour line is extracted from the sketch-like image in the second storage means by using the quantified value and the first threshold value. Means, and a shadow extraction means for inputting the second threshold value and generating a shadow portion in the sketch-like image in the fourth storage means by using the original image, the second threshold value and a shadow expression method; , An image displaying the sketch-like image in the second storage means Provides an environment to specify a display unit, a modification of the policy of the drawing style image,
Correction input means for outputting the correction information, and the correction information as an input to correct the sketch-like image in the second storage means in accordance with the correction information to generate the second
An image correction device for writing back to the storage device, and generating an image having a characteristic of rough drawing from the original image.