WO2004053796A1 - 類似画像抽出装置、類似画像抽出方法および類似画像抽出プログラム - Google Patents
類似画像抽出装置、類似画像抽出方法および類似画像抽出プログラム Download PDFInfo
- Publication number
- WO2004053796A1 WO2004053796A1 PCT/JP2003/015852 JP0315852W WO2004053796A1 WO 2004053796 A1 WO2004053796 A1 WO 2004053796A1 JP 0315852 W JP0315852 W JP 0315852W WO 2004053796 A1 WO2004053796 A1 WO 2004053796A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- image
- pixel
- value
- original image
- collation
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/97—Determining parameters from multiple pictures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
- G06V10/751—Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching
- G06V10/7515—Shifting the patterns to accommodate for positional errors
Definitions
- the present invention relates to a similar image extracting device, a similar image extracting method, and a similar image extracting program.
- the present invention has been made in view of the above problems, and has an image search accuracy of It is an object of the present invention to provide a similar image extraction device, a similar image extraction method, and a similar image extraction program that can perform processing at high cost and at high speed.
- the present invention uses some pixels in the original image to search for similar site candidates in the search target image, and compares these candidates with the original image for each pixel. A part similar to the original image is extracted from the search target image.
- the present invention includes an image data acquisition unit, a reference pixel gradation value acquisition unit, a comparison pixel gradation value acquisition unit, a comparison value calculation unit, and a scanning unit. With these means, the original image data and the search target image data are acquired, and while the search target image is scanned in an area of the same size as the original image, the gradation value of the collation pixel and the gradation value of the comparison pixel are obtained. A collation value indicating the similarity is calculated.
- the collation pixels are arranged at a plurality of positions in the original image, but have a smaller number of pixels than the number of pixels in the original image. Therefore, by scanning the search target image for each area of the same size as the original image, only a part of the pixels in the search target image and some pixels of the original image are compared, and the search target image is scanned. With, similar parts in the original image can be found as candidates. Therefore, the processing can be performed at a very high speed as compared with the processing for examining whether or not all the pixels of the search target image are similar to the original image.
- a region similar to the original image is extracted by the similar region extracting unit from the candidates found as described above. That is, since these candidates indicate that the similarity to the original image is high by the matching value indicating the similarity, any of these candidates is considered to be a search target. Therefore, after limiting the candidates as described above, the similarity between all the pixels in the candidate and all the pixels in the original image is calculated. As a result, it is highly probable that the most similar area is most similar to the original image, and this area is extracted. As a result, an image similar to the original image can be searched for in the search target image.
- the similarity is calculated for all the pixels of the original image and all the pixels in a plurality of regions which have been candidates by the above-described processing. Since only the pixels of the part are used, the candidate can be determined at high speed.Since this candidate is a part of the search target image, the target to calculate the similarity by comparing with all the pixels of the original image can be determined. Part of the search target image. Therefore, the processing can be performed at a very high speed as compared with the processing for examining whether or not all the pixels of the search target image are similar to the original image.
- the above-mentioned candidates are extracted based on a collation value indicating the similarity between the collation pixel and the comparison pixel. According to the determination based on the collation value, it is possible to determine whether the collation pixel is similar to the comparison pixel, but whether all pixels in the original image and all pixels in the candidate are very similar. It cannot be determined. However, if a configuration is adopted in which a plurality of the above candidates are detected based on the collation value, an area having a high degree of similarity to the original image is not ultimately extracted based on the collation value alone, but the most similar region is extracted from the plurality of candidates. Regions can be extracted.
- the search target site is included in these candidates with a very high probability. Therefore, even with a configuration in which candidates are listed with a small number of collation pixels and all pixels in the original image are compared, search can be performed with extremely high accuracy.
- the collation pixel is a part of the original image, and by comparing the collation pixel with a comparison pixel in the search target image, it is determined whether or not the compared portion is similar to the original image.
- a configuration in which the collation pixels are arranged every 5 to 10 pixels vertically and horizontally in the original image, and each pixel corresponding to 1 /, 8 to 1/16 of the number of vertical or horizontal pixels in the original image A configuration in which the collation pixels are arranged at the same time can be adopted. According to such a configuration, it is possible to set the reference pixel without causing positional deviation from the original image.
- various pitches other than the above-mentioned specific example can be adopted as the pitch of the matching pixels, and the pitch can be adjusted according to the size of the original image or the search target image, the required search accuracy, and the search speed. .
- the collation value only needs to indicate the similarity between the gradation value of the collation pixel and the gradation value of the comparison pixel according to the value.
- the collation pixel corresponding to the same position and the comparison pixel are compared.
- the similar region extracting means When extracting a region having a high degree of similarity from the collation value as a candidate, the similar region extracting means considers some of the higher-ranked regions having a small difference as candidates and considers a region whose difference is equal to or less than a predetermined threshold as a candidate.
- the candidate can be extracted only by the addition / subtraction operation of the gradation value between the collation pixel and the comparison pixel, and the candidate can be extracted at high speed.
- Various configurations can be employed for extracting a region having a high degree of similarity as a candidate from the collation value by the similar region extracting means.
- calculate the collation value It is possible to adopt a configuration in which the searched area and its matching value itself are stored in a storage medium, and after the scanning of the search target image is completed, an area having a high degree of similarity is extracted by referring to the storage medium.
- data indicating an area that gives a high similarity collation value to the storage medium may be sequentially stored up to a predetermined number of areas. .
- the collation value calculating means when the collation value is calculated by the collation value calculating means, the collation value in the already stored area is referred to, and when the calculated collation value indicates a higher similarity, the storage medium is stored. Update the stored contents.
- the data indicating the area stored in the storage medium at the time when the scanning is completed is data relating to some higher-order areas having a high degree of similarity. Therefore, the similar area extracting means can process the upper area having a high degree of similarity only by referring to the data in the storage medium, and at the same time as the end of scanning, the gradation values of almost all the pixels in the above area. The process can be shifted to a process of calculating the similarity between and the tone values of substantially all the pixels of the original image.
- the similarity may be information indicating the degree of similarity between all the pixels of the original image and all the pixels of each of the above-mentioned candidates, and may be the same as the above-mentioned collation value. It can be configured by a value obtained by adding the difference of the gradation value to all the pixels of each candidate.
- a part similar to the original image can be searched from the search target image by the processing by the similar region extraction means.
- the scanning means are sequentially changed to calculate the matching value. If each area is selected so that it does not overlap during the scanning process, fine adjustment of each area may result in higher similarity . Therefore, after extracting the area having a high degree of similarity by the similar area extracting means, an area having the same size as the original image is set in the vicinity of the area.
- the degree of similarity between the set area and the original image may be calculated. In such a configuration, a region having the highest ⁇ similarity among the similarities is set as a similar image of the original image, and a portion of the search target image that best matches the original image is regarded as a similar image. Can be extracted.
- the above-described similar image extracting apparatus may be implemented independently, or may be implemented together with another method while being incorporated in a certain device. Includes various embodiments, and can be appropriately changed.
- the method of searching for a similar image by calculating the similarity for all pixels of the original pixel after limiting the detection of the search region from some pixels of the original image described above is performed according to a predetermined procedure. It goes without saying that there is an invention in the procedure at the root of proceeding. Therefore, the present invention can be applied also as a method, and the invention according to claim 6 has basically the same operation.
- a predetermined program may be executed by the similar image extracting device. The present invention is also applicable as this program, and the invention according to claim 7 has basically the same operation.
- any storage medium can be used to provide the program.
- a magnetic recording medium or a magneto-optical recording medium may be used, and the same applies to any recording medium to be developed in the future.
- the concept of the present invention is not completely different, and it is necessary to record a part on a recording medium.
- a form in which the data is appropriately read in accordance with the information is also included.
- Figure 1 is a block diagram of the computer system.
- FIG. 2 is a functional block diagram of the similar image extraction program.
- FIG. 3 is a flowchart of a process performed by the similar image extraction program.
- FIG. 4 is an explanatory diagram illustrating a processing example according to the present embodiment.
- FIG. 5 is an explanatory diagram illustrating a processing example according to the present embodiment.
- FIG. 1 is a block diagram showing a computer system that executes a similar image extraction program according to an embodiment of the present invention.
- a computer system 10 has a scanner as an image input device.
- Each input device is capable of generating image data that represents an image by dot matrix pixels and outputting the image data to the computer main unit 12, where the image data is RGB. By displaying 256 gradations for each of the three primary colors, approximately 1.67 million colors can be expressed.
- a flexible disk drive 13a as an external auxiliary storage device, a hard disk 13b, and a CD-ROM drive 13c are connected to the computer body 12.
- the hard disk 13b stores the main programs related to the system, and stores necessary programs and image data from the flexible disk 13a1 and CD-ROM13c1 as needed. It is readable.
- a modem 14a is connected as a communication device for connecting the computer body 12 to an external network or the like, and is connected to the external network via the same public communication line or the like. It can be connected and downloaded software and image data can be introduced.
- the modem 14a accesses the outside through a telephone line.
- a keyboard 15a and a mouse 15b are connected to operate the computer body 12.
- the original image data and the search target image data in the present invention are respectively input to the computer main body 12 from the input device, the external auxiliary storage device, the communication device, and the like.
- a display 17a and a color printer 17b are provided as image output devices.
- the display 17a has a display area of 1024 pixels in the horizontal direction and 768 pixels in the vertical direction, and each pixel can display the above mentioned 1670,000 colors. I have. Of course, this resolution is merely an example, and can be changed as appropriate, such as 640 ⁇ 480 pixels or 800 ⁇ 600 pixels.
- the operating system (OS) 12a is operating as the basic program, and the operating system 12a has a display for displaying on the display 17a.
- a printer driver (PRTDRV) 12c that prints to the driver (DSPDRV) 12b and the color printer 17b is incorporated.
- These drivers 12b and 12c depend on the model of the display 17a and color printer 17b, and can be added to and changed from the operating system 12a according to each model. is there. Also, depending on the model, additional functions beyond the standard processing can be realized. In other words, it is possible to realize various additional processes within the allowable range while maintaining a common processing system on the standard system of the operating system 12a.
- the computer 12 includes a CPU 12 e, a ROM 12 f, a RAM I 2 g, and an I / O l 2 h.
- the CPU 12e that executes the arithmetic processing uses the ROM 12f as a temporary work area setting storage area or writes it to the RAM I 2g while using it as a program area. It executes the basic program as appropriate to control external devices and internal devices connected via I / O 12h.
- the application (APL) 12 d is executed on the operating system 12 a as the basic program.
- the processing contents of the application 12d are various, and the operation of the keyboard 15a or the mouse 15b as an operation device is monitored, and when operated, various external devices are appropriately applied.
- the processing result of the application 12 d is output as print data to the above-described color printer 17 b via the printer driver 12 c, and the color printer 17 b is a color printer.
- the color printer 17 b is a color printer.
- the similar image extraction program according to the present invention can be provided as the above application 12d, or as a program that realizes a part of the functions of the application 12d. Can also be. It is particularly preferable to incorporate the similar image extraction program according to the present invention into a panorama photograph synthesis program image search program.
- FIG. 2 shows the functional block diagram of the similar image extraction program and the data used in the processing.
- Figure 3 shows the processing flow of the similar image extraction program.
- the similar image extraction program 20 includes an image data acquisition module 21, a reference pixel gradation value acquisition module 22, a scanning unit 23, and a similar area extraction module 24.
- the scanning unit 23 further includes a comparison pixel gradation value acquisition module 23a and a collation value calculation module 23b.
- the original image data and the search target image data are stored on the hard disk 13 b.
- the data may be data obtained from a device such as 1b, or may be data obtained through a modem 14a or the like.
- the image data acquisition module 21 proceeds to step S100 to The original image data is read from the disk 13b and temporarily stored in RAM I 2g (original image data 12g 1).
- the following processing may be performed by acquiring all the color components of the image data, a specific color component may be acquired, or a component other than RGB (for example, a luminance component) may be acquired. Good.
- step S 105 the reference pixel gradation value acquisition module 22 sets a reference pixel for the original image data 12 g 1, and sets the reference pixel gradation value (reference pixel gradation value 12 g 3) Get.
- FIG. 4 shows an example of processing in the present embodiment, which will be described in detail later.
- the collation pixels are set at a fixed pixel pitch, and the collation pixels can be set for each lZm (m is a natural number) of the number of vertical and horizontal pixels of the original image, or can be set for each pixel (1 is a natural number).
- the position can be determined by various methods such as setting.
- step S110 the image data acquisition module 21 reads out the search target image data from the hard disk 13b and temporarily stores it in the RAM I 2g (search target image data 12g2).
- steps S115 to S140 a loop process is performed under the control of the scanning unit 23. That is, in step S115, the comparison pixel tone value acquisition module 23a sets an area having the same size as the original image on the search target image (hereinafter, referred to as a comparison area). Further, in step S120, the comparison pixel gradation value acquisition module 23a sets a comparison pixel corresponding to the same position as the above-mentioned comparison pixel in the comparison area, and sets the gradation value (comparison pixel gradation). Obtain the adjustment value 1 2 g 4).
- the collation value calculation module 23b calculates the collation value with reference to the collation pixel gradation value 12g3 and the comparison pixel gradation value 12g4.
- the collation value is the gradation value of the collation pixel and the comparison pixel corresponding to the same position. This value is the sum of the differences. The smaller the value, the higher the similarity. Further, the collation value calculation module 23b performs processing of storing the top three similarities indicated by the collation values in steps S130 and S135 in RAM I2g.
- RAM I 2 g has a storage capacity for storing a collation value and a collation value 1 2 g 5 indicating a position of an area to which the collation value is given, and a memory capable of storing three locations. Capacity is secured.
- step S130 it is determined whether or not the calculated collation value is smaller than the collation value already stored in the RAM 12g (high similarity) (step S130), and is smaller than the existing collation value. If it is determined that the collation value has been calculated, the collation value 12 g5 in the RAM I 2 g is updated with the collation value calculated above (step S135). That is, the largest matching value is deleted from the existing matching values 1 2 g 5 and the calculated matching value is stored in RAM 12 g. If it is not determined in step S130 that the collation value is smaller than the existing collation value, the update of RAM12g is skipped.
- step S140 it is determined whether or not the process of calculating the collation value for the comparison region has been completed for all the regions in the search target image, and the processes in step SI15 and thereafter are determined until it is determined that the process has been completed. repeat.
- step S115 a new comparison area is set so as not to overlap within the search target image. Of course, the comparison area is changed while partially overlapping, and the processing proceeds. You may.
- the matching value is calculated for the entire area in the search target image in step S 140. After it is determined that the processing to be performed is completed, the top three places of the similarity in the comparison value 12 g5 of the RAM 12 g are stored in each comparison area. For the first three iterations Since only three or less existing collation values are stored in the RAM 12g, the process of storing the first three locations without discrimination may be performed. Of course, the three maximum matching values may be stored as the initial values.
- step S145 the comparison area indicated by the comparison value 12g5 is set as a candidate, and the similar area extraction module 24 calculates the similarity between each of the candidates and the original image.
- a peripheral area having the same size as the original image is set at a position obtained by moving the position of each candidate up, down, left, and right, and the similarity between the image of the peripheral area and the original image is calculated.
- the similarity is a value obtained by adding the difference between the tone values of the pixels corresponding to the same t position for all the pixels belonging to the original image and the comparison region (or the peripheral region).
- the similarity in the present embodiment is the same as the above collation value, although the number of target pixels is different from the matching value, but the concept is the same, and the smaller the similarity, the smaller the original image and the comparison area (or the surrounding area). ) Are similar.
- the comparison area or the peripheral area having the minimum value (most similar) among the similarities calculated in step S145 is determined as a similar image, and the similar image is determined. Is stored in RAM 12 g (similar image data 12 g 6).
- FIGS. 4 and 5 are explanatory diagrams illustrating a state in which a part similar to the image indicated by the original image data 12 g1 is searched from the search target image data 12 g2 of a photograph in which a mountain is a subject.
- the reference pixel gradation value acquisition module 22 includes a plurality of reference pixels X in the original image data 12 g 1 as shown on the left side of FIG. Set X for each collation pixel. To obtain the gradation value of.
- a comparison area C is set for the search target image data 12g2, and the search target image is set. A process of sequentially changing the comparison area C in the image data 1 2 g 2 is performed.
- the comparison area C is moved from the left to the right, and when the comparison area C reaches the right end, the comparison area C is moved down and again from the left to the right, and the search target image data 1 2 g Scan for the whole of 2. Further, in this scanning process, the comparison pixel gradation value acquisition module 23a sets the comparison pixel X in the comparison area. Comparison pixel X corresponding to the position of. Set, acquires the gradation value of each ratio ⁇ element X e. The matching value calculation module 23b calculates the matching value by adding the difference between the matching pixel and the comparison pixel.
- i is a number for specifying a pixel position in the horizontal direction
- j is This is a number that specifies the pixel position in the vertical direction, and both are natural numbers.
- the origin in each parenthesis is the coordinates of the left end of the original image and the comparison area. Therefore, the comparison pixel in the original image and the comparison pixel in the comparison area are specified by the combination of i and j.
- T indicates the pitch of the matching pixel and the comparison pixel, and indicates that pixels are acquired at a predetermined pitch in the original image and the matching value is calculated. That is, reference pixel X.
- the collation value calculation module 23b stores the collation values of the top three in the order of the degree of similarity in RAM 12g.
- Similar region extraction module 2 4 matching value stored in the RAM 1 2 g, 1 2 g 5 refers to the identifies three the position of the comparison area indicated by the matching value 1 2 g 5.
- An area (peripheral area) of the same size as the comparison area is set on the upper, lower, left, and right sides of the comparison area to determine the similarity between all pixels of the original image and all pixels of the comparison area (or peripheral pixels). calculate. That is, in Figure 5 As shown in the figure, in the situation where the comparison areas C 1, C n, etc. are extracted as candidates based on the comparison value 12 g 5, the similarity is calculated by changing the position of the comparison area as shown by the broken line. I do. .
- the similarity is calculated according to the formula ⁇ (X. (i, j)- ⁇ ⁇ ⁇ (i, j)) shown in FIG.
- i and j are natural numbers and indicate the coordinates of the original image, the comparison area, and the surrounding area. Again, the origin is the coordinates of the left edge of the image.
- i and j have a larger value range than that shown in Fig. 4 above
- i is the value range of the number of horizontal pixels in the original image, the comparison region, and the peripheral region
- j is the value of the original image, the comparison region, and the peripheral region.
- the number of vertical pixels is the value range.
- n is a natural number, which is a number that specifies the comparison area and its surrounding area.
- a region most similar to the original image can be extracted from the comparison region and the peripheral region.
- the comparison region is a region extracted as a similar image candidate from the search target image data 12g2 based on the collation value shown in FIG. 4, and is therefore the most significant region based on the similarity shown in FIG.
- the region extracted as a similar part has a very high probability of being the part most similar to the original image in all of the search target image data 12 g 2. Therefore, a similar image can be determined by the above processing.
- the force S that determines the candidate for the comparison area based on only a part of the pixels, and the similar area extraction module 24 finally compares all the pixels.
- the comparison target is expanded to the surrounding area, a similar image can be searched very accurately.
- similar area extraction module 24 The process of calculating the similarity only for the comparison region as the candidate and extracting the region having the highest similarity may be performed without expanding the target of calculating the similarity between the surrounding regions. .
- the extracted region is not necessarily the most similar to the original image in the search target image, but a portion close to the most similar image can be searched very quickly.
- the scanning unit 23 does not necessarily have such a configuration that all the search target images are to be scanned while avoiding duplication when setting the comparison area. For example, when the collation value calculated by the collation value calculation module 23b is very large (when the similarity is very small), the scanning pitch is increased when the next comparison area is set. It is possible to adopt a configuration in which the comparison area is set farther and the scanning pitch is reduced as the collation value increases. Further, a configuration may be adopted in which the number of collation pixels is reduced when the collation value is extremely large. According to these configurations, redundant processing can be effectively reduced, and candidate comparison regions can be extracted more quickly.
- the present invention can be applied to various situations. For example, as described above, when it is desired to specify an area that matches between connected pictures when synthesizing a panoramic picture, the original image is cut out from one image, and the similar image extraction program according to the present invention is applied to the other image. By doing so, it is possible to easily identify a site that matches in both images. Therefore, it is possible to easily specify the position connecting the two. In this case, even if a plurality of candidates for the comparison area are extracted, a condition that a part regarded as a similar image must exist on the left side of one image and exist on the right side of the other image. For example, the processing can be completed faster. Industrial applicability
- a connection position search for connecting panorama photograph images, and the like a search target It is possible to provide a computer or the like that realizes a function of searching for a part similar to the original image in the image.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Databases & Information Systems (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Quality & Reliability (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Image Analysis (AREA)
Abstract
Description
Claims
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
EP03780724A EP1574991A4 (en) | 2002-12-11 | 2003-12-11 | EXTRACTION DEVICE FOR SIMILAR PICTURES, EXTRACTION PROCESSES FOR SIMILAR PICTURES AND EXTRACTION PROGRAM FOR SIMILAR IMAGES |
US10/540,787 US20060115148A1 (en) | 2002-12-11 | 2003-12-11 | Similar image extraction device, similar image extraction method, and similar image extraction program |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2002-358985 | 2002-12-11 | ||
JP2002358985A JP2004192307A (ja) | 2002-12-11 | 2002-12-11 | 類似画像抽出装置、類似画像抽出方法および類似画像抽出プログラム |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2004053796A1 true WO2004053796A1 (ja) | 2004-06-24 |
Family
ID=32500917
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2003/015852 WO2004053796A1 (ja) | 2002-12-11 | 2003-12-11 | 類似画像抽出装置、類似画像抽出方法および類似画像抽出プログラム |
Country Status (5)
Country | Link |
---|---|
US (1) | US20060115148A1 (ja) |
EP (1) | EP1574991A4 (ja) |
JP (1) | JP2004192307A (ja) |
CN (1) | CN1711558A (ja) |
WO (1) | WO2004053796A1 (ja) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7817861B2 (en) | 2006-11-03 | 2010-10-19 | Symantec Corporation | Detection of image spam |
Families Citing this family (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2003058554A1 (fr) * | 2001-12-28 | 2003-07-17 | Nikon Corporation | Dispositif de traitement d'image permettant d'effectuer un jugement de similitude entre des pixels et programme de traitement d'image |
JP4213112B2 (ja) * | 2004-11-09 | 2009-01-21 | シャープ株式会社 | 画像検索装置、画像形成装置、画像検索装置の制御方法、画像検索プログラムおよびコンピュータ読取り可能な記録媒体 |
JP4038777B2 (ja) * | 2006-02-28 | 2008-01-30 | ソニー株式会社 | 登録装置及び認証装置 |
AT504213B1 (de) * | 2006-09-22 | 2008-04-15 | Ipac Improve Process Analytics | Verfahren zum ähnlichkeitsvergleich von gegenständen |
CN101488129B (zh) * | 2008-01-14 | 2011-04-13 | 夏普株式会社 | 图像检索装置及图像检索方法 |
TWI395051B (zh) * | 2008-12-31 | 2013-05-01 | Altek Corp | Panorama Image Automatic Shooting Method for Digital Camera |
CN101859326B (zh) * | 2010-06-09 | 2012-04-18 | 南京大学 | 一种图像检索方法 |
EP2407928A1 (en) * | 2010-07-16 | 2012-01-18 | STMicroelectronics (Grenoble 2) SAS | Fidelity measurement of digital images |
JP2012203458A (ja) * | 2011-03-23 | 2012-10-22 | Fuji Xerox Co Ltd | 画像処理装置及びプログラム |
JP5919979B2 (ja) * | 2012-04-06 | 2016-05-18 | 富士ゼロックス株式会社 | 画像処理装置及び画像処理プログラム |
CN105100639B (zh) * | 2015-07-20 | 2017-11-24 | 京东方科技集团股份有限公司 | 一种显示方法及显示装置 |
CN105868788A (zh) * | 2016-04-01 | 2016-08-17 | 乐视控股(北京)有限公司 | 图像文件比对方法和装置 |
US10901674B2 (en) * | 2017-12-07 | 2021-01-26 | Honeywell International Inc. | Protocol for high integrity personal electronic device applications |
CN111046893B (zh) * | 2018-10-12 | 2024-02-02 | 富士通株式会社 | 图像相似性确定方法和装置、图像处理方法和装置 |
CN115100199B (zh) * | 2022-08-24 | 2022-11-22 | 宁波鑫芯微电子科技有限公司 | 一种晶圆低纹理缺陷的检测方法 |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPS59211167A (ja) * | 1984-05-04 | 1984-11-29 | Hitachi Ltd | パタ−ンの位置検出装置 |
JPH05159064A (ja) * | 1991-12-09 | 1993-06-25 | Nippon Telegr & Teleph Corp <Ntt> | 画像探索装置 |
JPH0830784A (ja) * | 1994-07-14 | 1996-02-02 | Sumitomo Metal Ind Ltd | 2値または濃淡値画像パターンマッチング装置 |
JPH08235357A (ja) * | 1995-02-24 | 1996-09-13 | Mitsubishi Electric Corp | 画像処理装置 |
JPH1021389A (ja) * | 1996-07-03 | 1998-01-23 | Mitsubishi Electric Corp | テンプレートマッチング方法およびその装置 |
Family Cites Families (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPS59226981A (ja) * | 1983-06-08 | 1984-12-20 | Fujitsu Ltd | パタ−ンマツチング方法および装置 |
US5530483A (en) * | 1994-10-11 | 1996-06-25 | Pixel Instruments Corp. | Delay detector apparatus and method for plural image sequences |
JPH09293082A (ja) * | 1996-04-26 | 1997-11-11 | Toshiba Corp | 画像検索装置及び画像検索方法 |
JPH11214485A (ja) * | 1998-01-26 | 1999-08-06 | Dainippon Screen Mfg Co Ltd | ウェハの回転方向決定方法 |
US7046842B2 (en) * | 1999-08-17 | 2006-05-16 | National Instruments Corporation | System and method for color characterization using fuzzy pixel classification with application in color matching and color match location |
-
2002
- 2002-12-11 JP JP2002358985A patent/JP2004192307A/ja not_active Withdrawn
-
2003
- 2003-12-11 CN CNA2003801030146A patent/CN1711558A/zh active Pending
- 2003-12-11 EP EP03780724A patent/EP1574991A4/en not_active Withdrawn
- 2003-12-11 US US10/540,787 patent/US20060115148A1/en not_active Abandoned
- 2003-12-11 WO PCT/JP2003/015852 patent/WO2004053796A1/ja not_active Application Discontinuation
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPS59211167A (ja) * | 1984-05-04 | 1984-11-29 | Hitachi Ltd | パタ−ンの位置検出装置 |
JPH05159064A (ja) * | 1991-12-09 | 1993-06-25 | Nippon Telegr & Teleph Corp <Ntt> | 画像探索装置 |
JPH0830784A (ja) * | 1994-07-14 | 1996-02-02 | Sumitomo Metal Ind Ltd | 2値または濃淡値画像パターンマッチング装置 |
JPH08235357A (ja) * | 1995-02-24 | 1996-09-13 | Mitsubishi Electric Corp | 画像処理装置 |
JPH1021389A (ja) * | 1996-07-03 | 1998-01-23 | Mitsubishi Electric Corp | テンプレートマッチング方法およびその装置 |
Non-Patent Citations (1)
Title |
---|
See also references of EP1574991A4 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7817861B2 (en) | 2006-11-03 | 2010-10-19 | Symantec Corporation | Detection of image spam |
Also Published As
Publication number | Publication date |
---|---|
EP1574991A4 (en) | 2006-09-27 |
US20060115148A1 (en) | 2006-06-01 |
JP2004192307A (ja) | 2004-07-08 |
CN1711558A (zh) | 2005-12-21 |
EP1574991A1 (en) | 2005-09-14 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2004053796A1 (ja) | 類似画像抽出装置、類似画像抽出方法および類似画像抽出プログラム | |
JP4338155B2 (ja) | 画像処理装置及びその方法、コンピュータ可読メモリ | |
WO2004051575A1 (ja) | 特徴領域抽出装置、特徴領域抽出方法および特徴領域抽出プログラム | |
US8446494B2 (en) | Automatic redeye detection based on redeye and facial metric values | |
JP4189506B2 (ja) | 画像処理のための装置、方法及び記録媒体 | |
JP2002152492A (ja) | 画像処理装置、方法及び記録媒体 | |
US20090231628A1 (en) | Image Processing Apparatus, Image Processing Method, Computer Program for Image Processing | |
JP2007004677A (ja) | コメントレイアウト装置、アルバム作成装置 | |
JP2011081803A (ja) | 赤目オブジェクト候補を分類する方法、コンピューター読み取り可能な媒体および画像処理装置 | |
JP2009237661A (ja) | 画像処理装置、画像処理方法、画像処理のためのコンピュータプログラム | |
JP2005151130A (ja) | 画像出力装置、画像出力方法ならびに記憶媒体、プログラム | |
JP2004054898A (ja) | 印刷データ生成装置、印刷データ生成方法およびそのプログラム | |
JP2006279460A (ja) | 画像処理装置、印刷装置、画像処理方法、および、画像処理プログラム | |
JP2007200091A (ja) | 画像処理装置 | |
JP4742642B2 (ja) | 画像処理装置、印刷装置、画像処理方法、および、画像処理プログラム | |
JP2008042291A (ja) | 印刷装置、および、印刷方法 | |
JP2003187257A (ja) | 画像作成システム | |
WO2022107635A1 (ja) | 撮影装置、画像の生産方法、および記録媒体 | |
JP4271428B2 (ja) | 文字認識方法および文字認識装置 | |
JP2008117345A (ja) | 印刷装置、画像検索方法、および、画像検索プログラム | |
JP4046509B2 (ja) | 画像生成方法および装置、画像処理方法および装置並びにプログラム | |
JP2003179737A (ja) | 画像処理装置、画像形成装置、および画像処理方法 | |
JP2006023949A (ja) | 画像処理装置及び画像処理方法 | |
JP2007253355A (ja) | 画像形成装置、画像形成システム及び情報処理装置 | |
JP4310943B2 (ja) | 文字認識装置、文字認識プログラム、文字認識方法 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AK | Designated states |
Kind code of ref document: A1 Designated state(s): CN US |
|
AL | Designated countries for regional patents |
Kind code of ref document: A1 Designated state(s): AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IT LU MC NL PT RO SE SI SK TR |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
WWE | Wipo information: entry into national phase |
Ref document number: 20038A30146 Country of ref document: CN |
|
ENP | Entry into the national phase |
Ref document number: 2006115148 Country of ref document: US Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 10540787 Country of ref document: US |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2003780724 Country of ref document: EP |
|
WWP | Wipo information: published in national office |
Ref document number: 2003780724 Country of ref document: EP |
|
WWP | Wipo information: published in national office |
Ref document number: 10540787 Country of ref document: US |
|
WWW | Wipo information: withdrawn in national office |
Ref document number: 2003780724 Country of ref document: EP |