WO1998020454A1 - Appareil d'extraction de structure - Google Patents
Appareil d'extraction de structure Download PDFInfo
- Publication number
- WO1998020454A1 WO1998020454A1 PCT/JP1996/003212 JP9603212W WO9820454A1 WO 1998020454 A1 WO1998020454 A1 WO 1998020454A1 JP 9603212 W JP9603212 W JP 9603212W WO 9820454 A1 WO9820454 A1 WO 9820454A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- pattern
- dimensional
- pattern data
- fourier
- free
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/42—Global feature extraction by analysis of the whole pattern, e.g. using frequency domain transformations or autocorrelation
- G06V10/431—Frequency domain transformation; Autocorrelation
Definitions
- an N-dimensional pattern for example, sound (one-dimensional), planar image (two-dimensional), three-dimensional (three-dimensional)] is collated based on spatial frequency characteristics, and a registered pattern and an illuminated pattern are compared.
- the present invention relates to a pattern extracting device for extracting a difference from a moving pattern and a moving pattern.
- the difference between two similar patterns has been extracted by a human visual check. That is, one pattern is set as a reference pattern, and the difference is extracted by comparing the reference pattern with the other pattern with human eyes. Extraction of common patterns (moving patterns) at different positions in two similar patterns (whole patterns) is also performed by human visual check. That is, one pattern is set as a reference pattern, and the reference pattern and the other pattern are compared with human eyes to extract a moving pattern.
- human visual checks can only deal with differences between two similar patterns or when the movement pattern is clear. In other words, if the entire pattern is composed of complex patterns, or if the differences or movement patterns are very small, it takes time to extract and accurate checks Can not do. In addition, it is difficult to detect all movement patterns on a platform where there are multiple movement patterns.
- An object of the present invention is to provide a pattern extracting device capable of extracting a difference between similar patterns and a moving pattern in a short time and accurately.
- the present invention provides a registered Fourier N-dimensional pattern data by performing an N-dimensional discrete Fourier transform on the N-dimensional pattern data of a registered pattern, and the N-dimensional pattern data of a collation pattern.
- N-dimensional discrete Fourier transform is applied to the data to create a matching Fourier N-dimensional pattern data, and the registered Fourier N-dimensional pattern data and the matching Fourier N-dimensional pattern
- the first pattern processing means performs N-dimensional discrete Fourier transform and N-dimensional discrete inverse Fourier on the combined free-dimensional N-dimensional pattern data obtained by combining them. Either one of the Fourier transforms is performed, the correlation peak in the correlation component area appearing in the synthesized Fourier N-dimensional pattern data subjected to the Fourier transform is obtained, and the obtained correlation peak is obtained.
- an N-dimensional discrete free-run is performed on the synthesized free-range data whose part is masked and a part of which is masked.
- the synthesized Fourier N-dimensional pattern data subjected to the linear transformation and the registered Fourier N-dimensional pattern data are re-synthesized, and the re-combined Fourier N-dimensional pattern data obtained as a result is recombined.
- N-dimensional discrete inverse Fourier transform is performed.
- the N-dimensional pattern data of the registered pattern is subjected to the N-dimensional discrete Fourier transform to generate the registered free N-dimensional pattern data, and the N-dimensional pattern data of the collation pattern is generated.
- the N-dimensional discrete Fourier transform is applied to the input data to create a matching Fourier N-dimensional pattern data.
- the registered Fourier N-dimensional pattern data and the matching Fourier N-dimensional pattern data are synthesized, and the first pattern is obtained from the synthesized Fourier processing N-dimensional pattern data obtained by this.
- the N-dimensional discrete Fourier transform or the N-dimensional discrete inverse Fourier transform is performed by the Fourier transform means, and the correlation component appearing in the synthesized Fourier N-dimensional pattern data subjected to the Fourier transform is obtained.
- the correlation peak in the end is obtained. Then, the periphery including and including the correlation peak is masked, and the synthesized Fourier N-dimensional pattern data in which a part of the correlation peak is masked is subjected to N-dimensional pattern processing by the first pattern processing means. If a discrete Fourier transform has been applied, an N-dimensional inverse discrete Fourier transform is applied, and the N-dimensional inverse discrete Fourier transform is applied in the first pattern processing means. If so, an N-dimensional discrete Fourier transform is performed.
- the synthesized Fourier N-dimensional pattern data subjected to the Fourier transform and the registered Fourier N-dimensional pattern data are re-synthesized, and the re-synthesized Fourier ⁇ N-dimensional N-dimensional discrete inverse Fourier transform is applied to the pattern data.
- the recombined hood subjected to the inverse Fourier transform obtained by this is In the N-dimensional pattern data, the outline of the difference is extracted, and the outline of the movement pattern is extracted. It is possible to know where the difference and the movement pattern are and where they exist.
- the present invention further combines the registered Fourier N-dimensional pattern data and the illuminated Fourier N-dimensional pattern data, and obtains an amplitude with respect to the Taisei Fourier N-dimensional pattern data obtained thereby.
- the second pattern processing means After performing suppression processing (1 og processing, T processing, etc.), one of ⁇ -dimensional discrete Fourier transform and ⁇ -dimensional discrete inverse Fourier transform is performed, and the second pattern processing means performs Performs amplitude restoration processing (inverse function processing such as 1 og processing and processing) on the synthesized Fourier ⁇ -dimensional pattern data that has been subjected to the Rier transform, and then synthesizes the data after the amplitude restoration processing has been performed.
- the free N-dimensional pattern data and the registered Fourier N-dimensional pattern data are re-synthesized, and the re-composed Fourier N-dimensional pattern data obtained by this is N-dimensional discrete inverse Fourier Perform the conversion.
- the present invention further synthesizes registered Fourier N-dimensional pattern data and collated Fourier N-dimensional pattern data, and converts the resulting data into Taisei Fourier N-dimensional pattern data.
- amplitude suppression processing (10 g processing and processing, etc.) on either of them
- one of N-dimensional discrete free-time transform and N-dimensional discrete inverse-free transform is performed, and the second pattern is applied.
- the reconstructed Fourier obtained by recombining the Fourier transformed N-dimensional pattern data and the registered Fourier N-dimensional pattern data subjected to the Fourier transform by the processing means. Performs N-dimensional discrete inverse Fourier transform on N-dimensional pattern data.
- the present invention further provides an N-dimensional discrete Fourier transform to the N-dimensional pattern data of the registered pattern, and then performs amplitude suppression processing (1 og processing, ⁇ processing, and the like), thereby obtaining a registered free space.
- N-dimensional discrete Fourier transform to the N-dimensional pattern data of the registered pattern, and then performs amplitude suppression processing (1 og processing, ⁇ processing, and the like), thereby obtaining a registered free space.
- the amplitude reconstruction processing (10 g processing and the inverse function processing such as "processing" is performed on the regenerated free-flowing N-dimensional pattern data.
- N-dimensional discrete inverse Fourier transform is performed.
- the present invention further provides an N-dimensional discrete file transformation of the N-dimensional pattern data of the registration pattern, and then performs an amplitude suppression process (1 og process, T process, etc.), thereby making the registration file available.
- the matching Fourier N-dimensional pattern data is created, and the combined Fourier N-dimensional pattern data subjected to Fourier transform by the second pattern processing means and the registered Fourier N-dimensional pattern data are re-synthesized. Then, an N-dimensional discrete inverse Fourier transform is applied to the recombined Fourier N-dimensional pattern data obtained by this.
- FIG. 1A to 1G are diagrams for explaining a process of extracting a difference between a registered pattern and a collation pattern in the pattern extraction device shown in FIG.
- FIG. 2 is a block diagram of a pattern extracting apparatus according to an embodiment of the present invention.
- FIG. 3 is a flowchart for explaining the registration (registration) pattern registration operation in the pattern extraction device shown in FIG.
- FIG. 4 is a flowchart for explaining an operation of extracting a difference between a registered pattern and a matching pattern in the pattern extracting apparatus shown in FIG.
- FIG. 5 is a flowchart, following FIG. 4, for explaining the operation of extracting the difference between the registered pattern and the matching pattern.
- 6A to 6G are diagrams for explaining a process of extracting a moving pattern in the pattern extraction device shown in FIG.
- FIG. 7 is a flowchart for explaining the movement pattern extraction operation in the pattern extraction device shown in FIG.
- FIG. 8 is a flowchart following FIG. 7 for explaining the movement pattern extraction operation. It is.
- FIG. 9 is a functional block diagram of the pattern extraction algorithm corresponding to the flowcharts shown in FIGS.
- FIG. 10 is a functional block diagram of the pattern extraction algorithm corresponding to the flowcharts shown in FIGS. 5. BEST MODE FOR CARRYING OUT THE INVENTION
- FIG. 2 is a block diagram of a pattern extracting apparatus according to an embodiment of the present invention, and illustrates a case where two-dimensional pattern data composed of image data is collated.
- 10 is an operation unit
- 20 is a control unit
- the operation unit 10 is a numeric keypad 10-1 and a display (LCD: Liquid Crystal Display) 10-2.
- CCD Charge Coupled Device
- the control unit 20 includes a control unit 20 — 1 having a CPU (Central Processing Unit), a ROM (Ead Only Memory) 20 — 2, and a RAM (Random Access Memory) 20 — 3, a disk (HD) 20-4, a frame memory (FM) 20-5, an external connection (I / F) 20-6, and a Fourier transform section (FFT) 20-7 and a pattern extraction program is stored in R 0 M 20 -2.
- a control unit 20 — 1 having a CPU (Central Processing Unit), a ROM (Ead Only Memory) 20 — 2, and a RAM (Random Access Memory) 20 — 3, a disk (HD) 20-4, a frame memory (FM) 20-5, an external connection (I / F) 20-6, and a Fourier transform section (FFT) 20-7 and a pattern extraction program is stored in R 0 M 20 -2.
- a control unit 20 — 1 having a CPU (Central Processing Unit), a ROM (Ead Only Memory) 20
- the reference pattern (registered pattern) is registered as shown in Fig. 3. That is, before starting the pattern matching, the user inputs the ID number assigned to the reference pattern using the numeric keypad 10-1 (step S301), and inputs the ID number to the CCD camera. Place the registration pattern at a predetermined position in the field of view of the camera 10-3. As a result, the original image of the registered pattern is converted to A
- control unit 20 It is given to the control unit 20 as a grayscale image (image data: two-dimensional pattern data) with 0 pixels and 256 gradations.
- the control unit 20-1 captures the image data of the registration pattern given by the operation unit 10 via the frame memory 20-5 (step S302), and captures the captured image data.
- the image data of the registered pattern (see Fig. 1A) is sent to the Fourier transform unit 20-7 to be subjected to a two-dimensional discrete Fourier transform (two-dimensional DFT: two-dimensional discrete Fourier transform). Step S304).
- the image data of the registered pattern shown in FIG. 1A becomes Fourier image data (registered Fourier image data) F A as shown in FIG. 1B.
- the control unit 20-1 receives the Fourier image data F A as the original image data of the registered pattern and inputs the data into the hard disk 20-4.
- the file is filed in correspondence with the 1D number (step S305).
- the two-dimensional discrete Fourier transform is described in, for example, “Introduction to Computer Image Processing, Japan Industrial Technology Center, pp. 44-45” (Reference 1).
- the difference between the registered pattern and the matching pattern is extracted as shown in FIG. That is, the user inputs the ID number assigned to the reference pattern using the numeric keypad 10-1 (step S401), and enters the field of view of the CCD camera 10-1. Place the matching pattern at a predetermined position in the range. As a result, the original image data of the matching pattern is
- grayscale image image data: two-dimensional pattern data
- control unit 20-1 receives the registration pattern filed in the disk 20-14. Then, the Fourier image data FA of the registered pattern corresponding to the ID number is read (step S).
- the control unit 20-1 captures the image data of the collation pattern provided from the operation unit 10 via the frame memory 20-5 (step S403), and captures the image data.
- the image data of the matching pattern (see Fig. 1C) is sent to the Fourier transform unit 20-7 to perform a two-dimensional discrete Fourier transform (two-dimensional DFT) (step S405).
- the image data of the collation pattern shown in FIG. 1C becomes Fourier image data (collation Fourier image data) FB as shown in FIG. 1D.
- the control unit 20-1 receives the Fourier image data FB of the collation pattern obtained in step S405 and the registered pattern read out in step S402.
- the free image data FA is synthesized (step S406) to obtain the synthesized free image data.
- the matching Fourier image data is assumed to be A e0
- the registered free image data is assumed to be B ⁇ . — Represented by 0).
- A, B, ⁇ , and ⁇ are also functions of frequency (four ⁇ ) space (u, V).
- a 'B' e J ( e — 0) A 'B' cos (0- ⁇ ) + j-A-B-sin ( ⁇
- Step S After obtaining the image data, perform amplitude suppression processing using the phase-only correlation method (Step S).
- 10 g processing is performed as the amplitude suppression processing. That is, an arithmetic expression of a synthetic full one Li E image data described above A 'B' e j - take the log of (0 3 ⁇ 4?), Log ( A 'B) - ej' ⁇ - in ⁇ and the child Thus, the amplitude ⁇ ⁇ ⁇ ⁇ ⁇ is suppressed to log (A ⁇ B) (A'B> log (A'B)).
- the above-mentioned phase-only correlation method is a cross-correlation modified so as to focus on a spatial phase change of an image, and is a synthetic Fourier in which amplitude information is suppressed and limited to only phase information. Image data is required.
- the influence of the illuminance difference between when the registered pattern is collected and when the collation pattern is collected is reduced. That is, by performing the amplitude suppression processing, the spectrum intensity of each pixel is suppressed, and there is no extraordinary value, so that more information is effective.
- 1 og processing is performed as amplitude suppression processing, but ⁇ processing may be performed. Further, the processing is not limited to the 10 g processing and the ⁇ processing, and any processing may be used as long as the amplitude can be suppressed.
- a constant value for all amplitudes with amplitude suppression For example, when the value is set to 1, that is, when only the phase is used, there is an advantage that the amount of calculation can be reduced and an amount of data can be reduced as compared with log processing or “processing”.
- control unit 20-1 After performing the amplitude suppression processing in step, the control unit 20-1 sends the synthesized Fourier image data subjected to the amplitude suppression processing to the Fourier transformation unit 20-7, and performs the second 2 A dimensional DFT is performed (step S408), whereby the synthesized free image data on which the amplitude suppression processing has been performed is a synthesized free image as shown in Fig. 1E. It becomes image data.
- the control unit 20-1 captures the synthesized Fourier image data obtained in step S408, and obtains a predetermined image including the central part from the synthesized Fourier image data.
- the pixel with the highest intensity (correlation peak) in the rear is extracted (step S409). In this case, a correlation peak appears near the center of the correlation component identifier.
- the control unit 20-1 masks the periphery including the correlation peak extracted in step S409 (step S410). That is, as shown in FIG. 1F, the area SO surrounded by the white dotted line is masked with respect to the combined free-flow image data shown in FIG. 1E. Then, the synthesized Fourier image data in which this region SO is masked is subjected to a two-dimensional discrete inverse Fourier transform (two-dimensional IDFT) (step S). 411), the amplitude restoration processing is performed on the synthesized Fourier image data on which the two-dimensional IDFT has been performed (step S412).
- the amplitude restoration process is a process in which the inverse function of the function performed in the amplitude suppression process in step S407 is performed on the amplitude. If 2 and 1 og e A, then e A.
- control unit 20-1 transmits the synthesized Fourier image data subjected to the amplitude restoration processing in step S412 and the registered file read out in step S402.
- the re-combined Fourier image data FA is recombined with the re-combined Fourier image data (step S4113), and the two-dimensional IDFT is performed on the recombined Fourier image data (step S413).
- S 4 14) to obtain recombined free-view image data as shown in FIG. 1G.
- the re-synthesis in step S413 means registration pattern B and collation pattern A.
- the contour of the pattern that exists only in the matching pattern appears at the corresponding position. .
- the pattern of the car is superimposed on a part of the matching pattern.
- the pattern of this car appears in Fig. 1G as a difference between the verification pattern and the registered pattern.
- the control unit 20-1 extracts the pattern of the vehicle appearing in FIG. 1G as a pattern existing only in the matching pattern (step S4 15) o
- the registered pattern does not have the car pattern superimposed thereon, and the car pattern is superimposed only on the matching pattern.
- the case has been described.
- the case where the car pattern is superimposed on both the registered pattern and the collation pattern, and the position of the car pattern is moving will be described below.
- control unit 20-1 performs the same processing as the flowchart shown in FIGS. 4 and 5, and extracts the movement pattern through the processing steps shown in FIG.
- FIG. 6E which is a process corresponding to FIG. 1E
- a correlation value P 1 indicating a background and a correlation value P 2 indicating a car appear.
- a mask is applied to the periphery including the correlation peak P 1 (see FIG. 6F).
- the control unit 20-1 performs a two-dimensional IDFT on the masked composite free image data (step S411), and performs the composite Fourier subjected to the two-dimensional IDFT.
- An amplitude restoration process is performed on the image data (step S412).
- the synthesized Fourier image data subjected to the amplitude restoration processing is read out in step S402.
- the registered registered free image data FA is recombined (step S4113), and a two-dimensional IDFT is performed on the recombined free image data obtained thereby (step S4).
- Step S4 14) obtains recombined Fourier image data as shown in FIG. 6G.
- the contours of the moving pattern existing in both the registered pattern and the matching pattern are represented by: Appears at the corresponding position in the matching pattern.
- the position of the vehicle in the registered pattern shown in FIG. 6A is moving.
- This car pattern appears in Fig. 6G as a moving pattern that exists at different positions in the matching pattern and the registered pattern.
- the control unit 20-1 extracts the pattern of the vehicle shown in FIG. 6G as a moving pattern existing at different positions of the registered pattern and the matching pattern (step). Step S 4 16).
- the two-dimensional IDFT is performed in the Fourier transform units 20-7, but may be performed in the CPU 20-1.
- the two-dimensional DFT is performed in step S408 shown in FIG. 4, but the two-dimensional DFT may be performed instead of the two-dimensional DFT. That is, instead of performing two-dimensional DFT on the synthesized Fourier image data on which the amplitude suppression processing has been performed, two-dimensional IDFT may be performed.
- two-dimensional IDFT may be performed in step D408.
- two-dimensional DFT is performed in step S411.
- the matching accuracy of the two-dimensional DFT and the two-dimensional IDFT does not change quantitatively.
- the two-dimensional IDFT is described in Reference 1 above.
- the 2-dimensional DFT in scan STEP S 4 0 8 may be subjected to amplitude suppression processing and then synthesized. That is, as shown in FIG. 7, the step S407 of FIG. 4 is eliminated, and the step S702 of reading out the registered Fourier image data FA and the image data of the collation pattern are eliminated. A step S703 for performing the amplitude suppression processing is provided between the step S704 for performing the input.
- a step S706 of applying two-dimensional DFT to the image data of the collation pattern and a collation Fourier after the amplitude suppression processing are combined with the image data FB and the registered Fourier image data FA.
- a step S707 for performing amplitude suppression processing is provided between the step S708 and the step S708.
- the combined Fourier image data subjected to the two-dimensional IDFT and the Fourier image data of the registered pattern read out in step 720 A step S714 for performing amplitude restoration processing is provided after the step S711 for recombining the signals.
- steps S701, S709 to S712, and S717 to S717 are the steps S401 and S410 shown in FIGS. 9 to S 4 12 and S 17 to S 4 17 are the same as those described above, and a description thereof will be omitted.
- the suppression rate of the amplitude of the synthesized free image data at this time is smaller than the case where the amplitude is suppressed after the synthesized free image data shown in FIGS. Therefore, the method of performing amplitude suppression processing on the synthesized Fourier image data shown in Fig. 4 is more effective than the method of performing amplitude suppression processing on the synthesized Fourier image data shown in Fig. 7 and generating the synthesized Fourier image data.
- the position of the difference and the matching accuracy of the movement pattern are improved. It should be noted that even when the amplitude suppression processing shown in FIGS. 7 and 8 is performed before the combined Fourier image data is obtained, two-dimensional! ) Instead of FT, two-dimensional IDFT may be performed.
- the two-dimensional pattern extraction processing has been described.
- the three-dimensional pattern extraction processing can be performed in the same manner, regardless of the two-dimensional or three-dimensional pattern.
- the multidimensional pattern extraction processing can be performed in the same manner.
- the amplitude suppression processing is performed, the amplitude suppression processing need not always be performed. Further, when all amplitudes are set to 1 in the amplitude suppression processing, that is, when the phase is limited, the amplitude restoration processing may not be performed.
- Fig. 9 shows a functional block diagram of the pattern extraction algorithm corresponding to the flow charts shown in Figs. 4 and 5, and Fig. 10 shows the flow chart shown in Figs. 7 and 8.
- a functional block diagram of the pattern extraction algorithm corresponding to Fig. 4 is shown.
- each functional block has the same step number as each step of the flowchart, and corresponds to the appended step number. The function of the step Each has.
- the registered pattern and the matching pattern are matched based on the spatial frequency characteristics, and the difference between similar patterns and the movement pattern are extracted as the matching result. This makes it possible to perform quality inspection, abnormality detection (analysis), and moving object detection in a short time and accurately.
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Description
Claims
Priority Applications (6)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/JP1996/003212 WO1998020454A1 (fr) | 1996-11-01 | 1996-11-01 | Appareil d'extraction de structure |
| US09/297,537 US6195460B1 (en) | 1996-11-01 | 1996-11-01 | Pattern extraction apparatus |
| DE69631845T DE69631845T2 (de) | 1996-11-01 | 1996-11-01 | Musterextrahierungsgerät |
| JP08536385A JP3035654B2 (ja) | 1996-11-01 | 1996-11-01 | パターン抽出装置 |
| EP96935527A EP1011073B1 (en) | 1996-11-01 | 1996-11-01 | Pattern extraction apparatus |
| KR1019997003884A KR100295586B1 (ko) | 1996-11-01 | 1999-04-30 | 패턴 추출장치 |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/JP1996/003212 WO1998020454A1 (fr) | 1996-11-01 | 1996-11-01 | Appareil d'extraction de structure |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO1998020454A1 true WO1998020454A1 (fr) | 1998-05-14 |
Family
ID=14154065
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/JP1996/003212 Ceased WO1998020454A1 (fr) | 1996-11-01 | 1996-11-01 | Appareil d'extraction de structure |
Country Status (6)
| Country | Link |
|---|---|
| US (1) | US6195460B1 (ja) |
| EP (1) | EP1011073B1 (ja) |
| JP (1) | JP3035654B2 (ja) |
| KR (1) | KR100295586B1 (ja) |
| DE (1) | DE69631845T2 (ja) |
| WO (1) | WO1998020454A1 (ja) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3468182B2 (ja) | 1999-12-21 | 2003-11-17 | 日本電気株式会社 | 画像マッチング装置および画像マッチング方法 |
| WO2008023726A1 (fr) * | 2006-08-22 | 2008-02-28 | Yamatake Corporation | Appareil de radar et procédé de mesure de la distance |
| JP2008089402A (ja) * | 2006-10-02 | 2008-04-17 | Konica Minolta Holdings Inc | 情報処理システム、プログラムおよび情報処理方法 |
| JP2010230578A (ja) * | 2009-03-27 | 2010-10-14 | Fujifilm Corp | 偏芯量測定方法 |
| KR101255952B1 (ko) | 2011-07-21 | 2013-04-23 | 주식회사 앤비젼 | 패턴층이 형성된 기판의 간섭 현상을 이용한 패턴검사방법 및 패턴검사장치 |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3846851B2 (ja) * | 2001-02-01 | 2006-11-15 | 松下電器産業株式会社 | 画像のマッチング処理方法及びその装置 |
| WO2005093654A2 (en) | 2004-03-25 | 2005-10-06 | Fatih Ozluturk | Method and apparatus to correct digital image blur due to motion of subject or imaging device |
| US10721405B2 (en) | 2004-03-25 | 2020-07-21 | Clear Imaging Research, Llc | Method and apparatus for implementing a digital graduated filter for an imaging apparatus |
| US9826159B2 (en) | 2004-03-25 | 2017-11-21 | Clear Imaging Research, Llc | Method and apparatus for implementing a digital graduated filter for an imaging apparatus |
| WO2005103610A1 (ja) * | 2004-04-22 | 2005-11-03 | The University Of Electro-Communications | 微小変位計測法及び装置 |
| US20060034531A1 (en) * | 2004-05-10 | 2006-02-16 | Seiko Epson Corporation | Block noise level evaluation method for compressed images and control method of imaging device utilizing the evaluation method |
| JP4525286B2 (ja) * | 2004-10-14 | 2010-08-18 | 沖電気工業株式会社 | 生体情報認証装置および認証方法 |
| JP4385139B2 (ja) | 2006-02-01 | 2009-12-16 | 国立大学法人電気通信大学 | 変位検出方法、及び、変位検出装置、変位検出プログラム、並びに、位相特異点マッチング処理方法、位相特異点マッチング処理プログラム |
| JP5630957B2 (ja) * | 2006-05-19 | 2014-11-26 | 日産化学工業株式会社 | ハイパーブランチポリマー及びその製造方法 |
| JP5178561B2 (ja) | 2009-02-06 | 2013-04-10 | Hoya株式会社 | パターン検査方法、パターン検査装置、フォトマスク製造方法、およびパターン転写方法 |
| CN102521590B (zh) * | 2011-04-29 | 2013-12-04 | 北京大学 | 一种基于方向的左右掌纹识别方法 |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH02206294A (ja) * | 1989-02-06 | 1990-08-16 | Nippon Telegr & Teleph Corp <Ntt> | 画像マッチング装置 |
| JPH0816785A (ja) * | 1994-06-27 | 1996-01-19 | Nec Corp | 画像認識装置および画像認識方法 |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| GB1570070A (en) * | 1976-01-07 | 1980-06-25 | Emi Ltd | Arrangement for correlating pictures |
| US5619596A (en) * | 1993-10-06 | 1997-04-08 | Seiko Instruments Inc. | Method and apparatus for optical pattern recognition |
| JPH07108526A (ja) | 1993-10-12 | 1995-04-25 | Nippon Steel Corp | スライシングマシン |
| US5684890A (en) * | 1994-02-28 | 1997-11-04 | Nec Corporation | Three-dimensional reference image segmenting method and apparatus |
| JPH07254062A (ja) | 1994-03-14 | 1995-10-03 | Yamatake Honeywell Co Ltd | 指紋照合装置 |
| US5915034A (en) * | 1995-05-02 | 1999-06-22 | Yamatake-Honeywell, Co., Ltd. | Pattern collation apparatus based on spatial frequency characteristics |
| US5886744A (en) * | 1995-09-08 | 1999-03-23 | Intel Corporation | Method and apparatus for filtering jitter from motion estimation video data |
| JPH09147115A (ja) * | 1995-11-20 | 1997-06-06 | Hamamatsu Photonics Kk | 人物照合装置 |
| US6067367A (en) * | 1996-10-31 | 2000-05-23 | Yamatake-Honeywell Co., Ltd. | Moving direction measuring device and tracking apparatus |
-
1996
- 1996-11-01 JP JP08536385A patent/JP3035654B2/ja not_active Expired - Fee Related
- 1996-11-01 EP EP96935527A patent/EP1011073B1/en not_active Expired - Lifetime
- 1996-11-01 US US09/297,537 patent/US6195460B1/en not_active Expired - Fee Related
- 1996-11-01 DE DE69631845T patent/DE69631845T2/de not_active Expired - Fee Related
- 1996-11-01 WO PCT/JP1996/003212 patent/WO1998020454A1/ja not_active Ceased
-
1999
- 1999-04-30 KR KR1019997003884A patent/KR100295586B1/ko not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH02206294A (ja) * | 1989-02-06 | 1990-08-16 | Nippon Telegr & Teleph Corp <Ntt> | 画像マッチング装置 |
| JPH0816785A (ja) * | 1994-06-27 | 1996-01-19 | Nec Corp | 画像認識装置および画像認識方法 |
Non-Patent Citations (1)
| Title |
|---|
| See also references of EP1011073A4 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3468182B2 (ja) | 1999-12-21 | 2003-11-17 | 日本電気株式会社 | 画像マッチング装置および画像マッチング方法 |
| WO2008023726A1 (fr) * | 2006-08-22 | 2008-02-28 | Yamatake Corporation | Appareil de radar et procédé de mesure de la distance |
| JP2008089402A (ja) * | 2006-10-02 | 2008-04-17 | Konica Minolta Holdings Inc | 情報処理システム、プログラムおよび情報処理方法 |
| JP2010230578A (ja) * | 2009-03-27 | 2010-10-14 | Fujifilm Corp | 偏芯量測定方法 |
| KR101255952B1 (ko) | 2011-07-21 | 2013-04-23 | 주식회사 앤비젼 | 패턴층이 형성된 기판의 간섭 현상을 이용한 패턴검사방법 및 패턴검사장치 |
Also Published As
| Publication number | Publication date |
|---|---|
| DE69631845T2 (de) | 2005-01-05 |
| EP1011073A1 (en) | 2000-06-21 |
| JP3035654B2 (ja) | 2000-04-24 |
| DE69631845D1 (de) | 2004-04-15 |
| KR20000053002A (ko) | 2000-08-25 |
| EP1011073B1 (en) | 2004-03-10 |
| EP1011073A4 (en) | 2000-06-21 |
| KR100295586B1 (ko) | 2001-08-07 |
| US6195460B1 (en) | 2001-02-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO1998020454A1 (fr) | Appareil d'extraction de structure | |
| CN101061489B (zh) | 用来处理视频数据的装置和方法 | |
| JPWO1998020454A1 (ja) | パターン抽出装置 | |
| Uddin et al. | Deepfake face detection via multi-level discrete wavelet transform and vision transformer | |
| Cox | Template matching and measures of match in image processing | |
| JP2021120814A (ja) | 学習プログラム、学習方法、及び情報処理装置 | |
| Wong et al. | Local statistics for generative image detection | |
| WO2019123917A1 (ja) | 画像照合装置 | |
| JPH08248045A (ja) | 速度測定方法および速度測定装置 | |
| Chen et al. | Stereoscopic ranging by matching image modulations | |
| CN119700062A (zh) | 一种非接触心率检测方法、系统及相关设备 | |
| CN112807000A (zh) | 鲁棒性脑电信号的生成方法及装置 | |
| US12425792B2 (en) | Video processing device and method | |
| Elbouz et al. | Correlation based efficient face recognition and color change detection | |
| JP2008225970A (ja) | 画像合成装置およびそれを用いた画像照合装置ならびに画像合成方法 | |
| Popescu et al. | VIBROTOOL—Software tool for change detection and diagnosis in vibration signals | |
| JPH10134196A (ja) | 距離測定装置 | |
| Bhowmik | Improved single-sensor-based modal identification using singular spectrum analysis | |
| Sitanggang et al. | Image Edge Detection for Batak Ulos Motif Recognition using Canny Operators | |
| JPH11296679A (ja) | 画像特徴抽出方法および装置並びに画像特徴抽出プロ グラムを記録したコンピュータ読み取り可能な記録媒 体 | |
| Deravi | Biometrics standards | |
| JP3197193B2 (ja) | 2次元信号フィルタリング装置 | |
| Kumar et al. | Image disparity estimation based on fractional dual-tree complex wavelet transform: A multi-scale approach | |
| CN121837127A (zh) | 一种基于多尺度边缘特征融合的骨折智能检测方法 | |
| JP3058951B2 (ja) | 画像パターン認識装置 |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| WWE | Wipo information: entry into national phase |
Ref document number: 96180491.2 Country of ref document: CN |
|
| AK | Designated states |
Kind code of ref document: A1 Designated state(s): CN JP KR US |
|
| AL | Designated countries for regional patents |
Kind code of ref document: A1 Designated state(s): AT BE CH DE DK ES FI FR GB GR IE IT LU MC NL PT SE |
|
| DFPE | Request for preliminary examination filed prior to expiration of 19th month from priority date (pct application filed before 20040101) | ||
| 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: 09297537 Country of ref document: US Ref document number: 1019997003884 Country of ref document: KR |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 1996935527 Country of ref document: EP |
|
| WWP | Wipo information: published in national office |
Ref document number: 1996935527 Country of ref document: EP |
|
| WWP | Wipo information: published in national office |
Ref document number: 1019997003884 Country of ref document: KR |
|
| WWG | Wipo information: grant in national office |
Ref document number: 1019997003884 Country of ref document: KR |
|
| WWG | Wipo information: grant in national office |
Ref document number: 1996935527 Country of ref document: EP |