CN1567370A - video processing method and device - Google Patents
video processing method and device Download PDFInfo
- Publication number
- CN1567370A CN1567370A CN 03146744 CN03146744A CN1567370A CN 1567370 A CN1567370 A CN 1567370A CN 03146744 CN03146744 CN 03146744 CN 03146744 A CN03146744 A CN 03146744A CN 1567370 A CN1567370 A CN 1567370A
- Authority
- CN
- China
- Prior art keywords
- image
- face
- analytic approach
- treatment method
- feature
- 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.)
- Pending
Links
Images
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
This invention discloses a kind of image processing method and apparatus. It classifies the captured face image by their features and stores them to individual group. The image in different group can be further classified according to the image quality. Then the image with better quality can be kept. It not only makes the use of memory body more effective, but also in favor of the subsequent image identification. This invention can help user to remember other's faces, and additionally, it can realize the purpose of automatic identification by using an acceptable means in sociality.
Description
Technical field
The present invention relates to a kind of method of handling image, refer to a kind of image treatment method and device especially.
Background technology
Because common people have the demand that is identified (recognized) psychologically, therefore, the sort of happiness when the other side tells our name, being commonly referred to as " sir " or " Miss " absolutely not can compare.In addition, can remember each contacted face and correctly say their name or title, often help to branch out and keep good interpersonal relation.
With regard to present developing technology, the design of intelligent memory and treating apparatus need comprise cross-cutting professional knowledges such as Neuscience, psychophysics and computer science.Usually people can rely on complicated " association " neural network and remember image of face, for example, and " he except not having beard, the math teacher of look the same my senior middle school ".With regard to the viewpoint that develops, this is the instinct that people have in order to seek survival, just so can reduce the probability of making a mistake.Yet if do not strengthen " association " by upgrading memory, memory will disappear.
For the relevance of emphasizing name and face with convenient memory, we are often by the feature of exaggerative face, allow the image of face can be distinct more, easier memory simultaneously, so individual skill can be by teaching, but effect is unknown.Another improves face's memory style and draws the face profile for using electronic product, and the affix relevent information is done in the future and searched.The personal digital assistant that many markets sell (PDA) provide such function.But this mode utmost point of dealing with problems lacks efficient, has often spent many times by after choosing various face features in the default feature of formula, but has to a neither face of shortage identification again naturally.Therefore, the best way keeps real image of face exactly and is aided with relevent information.But for the stranger takes pictures, not only rude in general occasion, the while is technical also has any problem, for example, and in the time of can't vacating both hands.
Summary of the invention
But the object of the present invention is to provide a kind of moment to obtain image treatment method and device that image of face is also handled efficiently.
In order to achieve the above object, the technology of the present invention solution is:
A kind of image processor comprises:
Face's arrangement for detecting is in order to detect an image, to find out image of face wherein;
Face's sorter is in order to be categorized as group according to face feature with this image of face;
The attribute classification device, the image of face that will be stored in each group in order to the quality according to image is classified;
Memory storage is in order to store this image of face according to group.
Face's arrangement for detecting of described image processor is found out this image of face by the detecting of class neural network analytic approach.
Face's arrangement for detecting of described image processor is found out this image of face by essential element analytic approach or the detecting of feature model analytic approach.
Face's sorter of described image processor is by essential element analytic approach this image of face of classifying.
Face's sorter of described image processor is according to face feature this image of face of classifying.
Face's sorter of described image processor is according to non-face feature this image of face of classifying, and this non-face feature comprises at least one in hair style, height, profile, clothes color, the hair color.
The attribute classification device of described image processor is by the classify image of face of this group of statistical analysis method.
The attribute classification device of described image processor is by ash classify image of face of this group of analytic approach all.
Described image processor comprises that also one stops judgment means, and whether it enters battery saving mode according to a preset signals or situation decision.
Described image processor also comprises a communication interface, and it is in order to access one external data bank.
A kind of image treatment method comprises the following steps:
A) in the image that obtains, find out an image of face;
B) this image of face is classified according to its feature, and be stored in other group;
C) image of face that will be stored in these indivedual groups is classified according to image quality;
D) data that will be stored in this group is exported.
The step a) of described image treatment method is finished by class neural network analytic approach.
The step a) of described image treatment method is finished by essential element analytic approach or feature model analytic approach.
The step b) of described image treatment method is finished by the essential element analytic approach.
The step b) of described image treatment method is according to face feature this image of face of classifying.
The step b) of described image treatment method is according to non-face feature this image of face of classifying, and this non-face feature comprises at least one in hair style, height, profile, clothes color, the hair color.
The step c) of described image treatment method is finished by statistical analysis method.
The step c) of described image treatment method by ash all analytic approach finish.
The step d) of described image treatment method is that this data is exported in the data storage device.
After adopting such scheme, the present invention not only can help the user to remember face efficiently, the more important thing is, the present invention use a kind of in the social life acceptable means reach the purpose of automatic identification.
Description of drawings
Fig. 1 is for using the typical situation of apparatus of the present invention;
Fig. 2 A is the front view of apparatus of the present invention preferred embodiment;
Fig. 2 B is the side view of apparatus of the present invention preferred embodiment;
Fig. 3 is the functional block diagram of preferred embodiment of the present invention;
Fig. 4 is the processing flow chart of preferred embodiment of the present invention.
Embodiment
Fig. 1 is for using the typical situation of apparatus of the present invention.This device is clipped in the shirt pocket place in camera lens mode forward, after starting by suitable mode, just can initiatively detects and store effective image of face, surpass preset value up to standby time, that is do not find target again, just decommission.With regard to the resolution of general camera lens (non-wide-angle lens or dolly camera) and 320240 pixels, the distance that obtains effective image of face is about 50 centimeters to 3 meters.For being applied to different occasions, for example long or wider coverage, the also replaceable suitable camera lens of this device.In addition, this device still can be selected different appearance designs, locates so that be hidden in cap, knot, glasses etc.Except can carrying, the present invention also can embed in motionless object of a stationkeeping or the device.For example, this device is located in the sales counter,, can be recognized automatically or assist the employee to recognize the regular guest, and reach safety or strengthen the purpose of service with this Auto-Sensing client face.
Fig. 2 A and Fig. 2 B are the front view and the side view of apparatus of the present invention preferred embodiment, and it is a pouch-type image processor, and called after " face's interceptor (face grabber) " 20.The external form of this face's interceptor 20 is just like a pen that volume is bigger, and is therefore very latent close when wearing or operate.Use a pinhole lens simultaneously, find even closely also be difficult to.Camera lens 21 is arranged in the rotation cap 22, and height that therefore can the person of being used adjusts.24 of pocket clip can be fixed in face's interceptor 20 shirt pocket or the place of any handled easily.Face's interceptor 20 has the LCD (LCD) 25 that can browse image of face and other information, the microphone 23 that is used to record and control knob 26.
Fig. 3 is the functional block diagram 30 of face's interceptor 20.The image sensor 31 that is used to obtain digitized video can be charge coupled cell (CCD) or complementary metal oxide semiconductor (CMOS).In order to save electric power, face's interceptor 20 uses battery saving mode usually, " wakes up " up to being activated device 33.Starter gear 33 can comprise the Digiplex that uses bluetooth (bluetooth) Radio Transmission Technology, control knob 34, infrared inductor, motion sensor etc.The most important thing is can start face's interceptor 20 by least eye-catching mode for starter gear 33.Central processing unit (CPU) 32 mainly provides the computing function of face's detecting and image compression.But memory body 35 comprises required impact damper (buffer) when calculating and the storage device of permanent storage image of face and other data.Typical display 36 is LCD (LCD), is mainly used in image of face and other data browsed.Communication interface 37 can comprise following: based on the wireless transmission of bluetooth science and technology, general-purpose serial bus (USB), infrared interface etc.Communication interface 37 can be obtained data by other device, perhaps data transmission is installed to other.For example, synchronous processing image of face and contact information between a computer and face's interceptor 20.Control knob 34 is used for following basic operation: set date and time, start the image classification pattern, start face's identification pattern, start exposal model, start recording mode, browse information in every way, remove or the correction data, start and the communication function of another device etc. with various patterns.
Fig. 4 is the preferred embodiment of treatment scheme of the present invention.Setting up procedure can be reached in every way by the described starter gear of leading portion.Once you begin, flow process just enters the loop of " detecting-classification-detecting-classification ... ", that is repeating step stops signal or meets default end condition up to receiving a) to step c).
Usually the actual state user pressed control knob 26 earlier and starts face's interceptor 20 before entering into the crowd, then, shook hands with many people in regular turn or to and fro.At this moment, image display device (digital camera) will be triggered and at step a1) produce an image.In the present invention, the form of image data can be digital pictures (digitalpicture), digital image (digital video), analogy image (analog video), image shelves (image files) without limits.At step a2) in, obtained image is then found out face area by face's arrangement for detecting.This face's arrangement for detecting can design according to algorithms such as feature model, neural network, and this is a prior art, so repeat no more.
Step b) mainly is with step a2 according to face feature) in the image of face that detects classified, and the group (cluster) that sets up the individual in impact damper can all be had people's portrait album one by one just as each by the person of taking pictures for storing its image data.Usually this step is to finish by essential element analytic approach (PCA), and it is a prior art, need not describe in detail.Except face feature, non-face feature also can be used as the auxiliary foundation of classification, for example clothing color, hair color, hair style, height, profile etc.
Step c) is then classified the image of face in each group according to image quality again.The technology of using is statistical method, for example by ash analytic approach (histogram analysis) sharpness of coming face features such as evaluate ocular, nose, mouth all.The technology of this kind image quality classification also is known, and the technology of the automatic focusing of very similar video camera.
By above-mentioned step b) and c) as can be known, the inventive method has two notable features: the image data amount that 1) can effectively control same people, because step b) is according to personal characteristics and it is classified in separately the group, and the capacity of each group can be limited, this situation may betide when shaking hands with the people back and forth, owing to repeating to take pictures through same people; 2) can avoid accumulating the image of too much inferior quality, because step c) is to do classification according to image quality, then the image of inferior quality just can be deleted.
Above-mentioned loop can stop when following predetermined condition takes place: surpass the predetermined operations period, outnumber preset value etc. above the image of face of accumulating in default empty lots (that is can not find any face) or the impact damper; Above-mentioned loop also can stop when receiving array apparatus to send the termination signal: motion sensor, Digiplex, infrared inductor, button etc.Usually, when surpassing default empty lots, for example can not find any face in 30 seconds, program can stop and switch to battery saving mode automatically.
Step d) is promptly carried out after stopping in above-mentioned loop.In step d), the image of performance optimal and the related data in the impact damper can be exported to display, inside or outside data bank or printing device etc.
Because the inventive method is record image of face when program start only, therefore required storage volume is very little, even is equipped with common low cost, for example 64,000,000 fast flash memory bank, just can store thousands of image of face easily, and still have living space other information of record and sound.
In addition, apparatus of the present invention still can switch to " identification pattern " according to circumstances.Under this pattern, step b ') step b) and c have been replaced).Step b ') can recognize the image of face of being detected by essential element analytic approach (PCA).Result's (for example name that combines with this image of face in the data bank) of identification then can be sent to wireless headset via communication interface 37.Therefore, the present invention not only helps the user to remember face efficiently, the more important thing is, the present invention use a kind of in the social life acceptable means reach the purpose of automatic identification.
The above only is the preferred embodiment of the inventive method and device, and is therefore all according to principle of the present invention and spirit modification person in addition, all should belong to scope of the present invention.
Claims (19)
1, a kind of image processor is characterized in that this device comprises:
Face's arrangement for detecting is in order to detect an image, to find out image of face wherein;
Face's sorter is in order to be categorized as group according to face feature with this image of face;
The attribute classification device, the image of face that will be stored in each group in order to the quality according to image is classified;
Memory storage is in order to store this image of face according to group.
2, image processor as claimed in claim 1 is characterized in that: this face's arrangement for detecting is found out this image of face by the detecting of class neural network analytic approach.
3, image processor as claimed in claim 1 is characterized in that: this face's arrangement for detecting is found out this image of face by essential element analytic approach or the detecting of feature model analytic approach.
4, image processor as claimed in claim 1 is characterized in that: this face's sorter is by essential element analytic approach this image of face of classifying.
5, image processor as claimed in claim 1 is characterized in that: this face's sorter is according to face feature this image of face of classifying.
6, image processor as claimed in claim 1 is characterized in that: this face's sorter is according to non-face feature this image of face of classifying, and this non-face feature comprises at least one in hair style, height, profile, clothes color, the hair color.
7, image processor as claimed in claim 1 is characterized in that: this attribute classification device is by the classify image of face of this group of statistical analysis method.
8, image processor as claimed in claim 1 is characterized in that: this attribute classification device is by ash classify image of face of this group of analytic approach all.
9, image processor as claimed in claim 1 is characterized in that: this device comprises that also one stops judgment means, and whether it enters battery saving mode according to a preset signals or situation decision.
10, image processor as claimed in claim 1 is characterized in that: this device also comprises a communication interface, and it is in order to access one external data bank.
11, a kind of image treatment method is characterized in that this method comprises the following steps:
A) in the image that obtains, find out an image of face;
B) this image of face is classified according to its feature, and be stored in other group;
C) image of face that will be stored in these indivedual groups is classified according to image quality;
D) data that will be stored in this group is exported.
12, image treatment method as claimed in claim 11 is characterized in that: this step a) is finished by class neural network analytic approach.
13, image treatment method as claimed in claim 11 is characterized in that: this step a) is finished by essential element analytic approach or feature model analytic approach.
14, image treatment method as claimed in claim 11 is characterized in that: this step b) is finished by the essential element analytic approach.
15, image treatment method as claimed in claim 11 is characterized in that: this step b) is according to face feature this image of face of classifying.
16, image treatment method as claimed in claim 11 is characterized in that: this step b) is according to non-face feature this image of face of classifying, and this non-face feature comprises at least one in hair style, height, profile, clothes color, the hair color.
17, image treatment method as claimed in claim 11 is characterized in that: this step c) is finished by statistical analysis method.
18, image treatment method as claimed in claim 11 is characterized in that: this step c) by ash all analytic approach finish.
19, image treatment method as claimed in claim 11 is characterized in that: this step d) is that this data is exported in the data storage device.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 03146744 CN1567370A (en) | 2003-06-23 | 2003-06-23 | video processing method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 03146744 CN1567370A (en) | 2003-06-23 | 2003-06-23 | video processing method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN1567370A true CN1567370A (en) | 2005-01-19 |
Family
ID=34471844
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 03146744 Pending CN1567370A (en) | 2003-06-23 | 2003-06-23 | video processing method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN1567370A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101594462B (en) * | 2009-05-27 | 2010-09-01 | 常州市盛和电子有限公司 | Explosion-proof dust-ignition-proof spherical camera |
| CN101527040B (en) * | 2008-03-05 | 2012-12-19 | 华为终端有限公司 | Method and system for processing images |
-
2003
- 2003-06-23 CN CN 03146744 patent/CN1567370A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101527040B (en) * | 2008-03-05 | 2012-12-19 | 华为终端有限公司 | Method and system for processing images |
| US8416314B2 (en) | 2008-03-05 | 2013-04-09 | Huawei Device Co., Ltd. | Method and system for processing images |
| CN101594462B (en) * | 2009-05-27 | 2010-09-01 | 常州市盛和电子有限公司 | Explosion-proof dust-ignition-proof spherical camera |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10776611B2 (en) | Method and system for identifying an individual in a digital image using location meta-tags | |
| US10853690B2 (en) | Method and system for attaching a metatag to a digital image | |
| US8369570B2 (en) | Method and system for tagging an image of an individual in a plurality of photos | |
| Costache et al. | In-camera person-indexing of digital images | |
| US8170343B2 (en) | Method and system for searching images with figures and recording medium storing metadata of image | |
| EP2336949B1 (en) | Apparatus and method for registering plurality of facial images for face recognition | |
| US20100014719A1 (en) | Storage apparatus and method, program, and playback apparatus and method | |
| CN101980242A (en) | Human face discrimination method and system and public safety system | |
| KR101252843B1 (en) | Method of Managing Image in a Mobile Communication Terminal | |
| CN105404878A (en) | Photo classification method and apparatus | |
| CN112199530A (en) | Multi-dimensional face library picture automatic updating method, system, equipment and medium | |
| CN103745223A (en) | Face detection method and apparatus | |
| Davis et al. | Using context and similarity for face and location identification | |
| WO2015131571A1 (en) | Method and terminal for implementing image sequencing | |
| CN104252618B (en) | method and system for improving photo return speed | |
| CN111723733A (en) | Highway toll collector smile rate statistical method and system | |
| CN1567370A (en) | video processing method and device | |
| CN103810294A (en) | Management method and intelligent terminal for multi-media data files | |
| Almansour et al. | I-privacy photo: Face recognition and filtering | |
| Singh et al. | Generic action recognition from egocentric videos | |
| TWI229823B (en) | Method and apparatus for processing image | |
| CN117312604A (en) | Face sample library establishment method and device, electronic equipment and storage medium | |
| CN109086464A (en) | A kind of picture searching system based on recognition of face | |
| CN117708056A (en) | Picture display method, device, chip, electronic equipment and medium | |
| CN114581968A (en) | Face big data detection and comparison system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |