CN106954002B - A kind of compression method and device of finger print data - Google Patents
A kind of compression method and device of finger print data Download PDFInfo
- Publication number
- CN106954002B CN106954002B CN201610011163.9A CN201610011163A CN106954002B CN 106954002 B CN106954002 B CN 106954002B CN 201610011163 A CN201610011163 A CN 201610011163A CN 106954002 B CN106954002 B CN 106954002B
- Authority
- CN
- China
- Prior art keywords
- finger print
- print data
- image data
- compression
- characteristic value
- 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.)
- Active
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
- H04N1/41—Bandwidth or redundancy reduction
- H04N1/411—Bandwidth or redundancy reduction for the transmission or storage or reproduction of two-tone pictures, e.g. black and white pictures
- H04N1/413—Systems or arrangements allowing the picture to be reproduced without loss or modification of picture-information
- H04N1/417—Systems or arrangements allowing the picture to be reproduced without loss or modification of picture-information using predictive or differential encoding
Abstract
The present invention relates to fingerprint identification technology fields, disclose the compression method and device of a kind of finger print data.In the present invention, the compression based on comentropy is carried out to source finger print data, is stored or is exported using compressed image data as collected finger print data to next processing unit.Under the premise of not influencing system performance, source data is compressed, reaching reduces bandwidth, memory space and the requirement for improving safety.Moreover, the algorithm based on comentropy is compressed based on feature decomposition (Eigen decomposition), make that it is more easier the realization of hardware and cost is lower.
Description
Technical field
The present invention relates to fingerprint identification technology field, in particular to a kind of compression method of finger print data.
Background technique
With the development of science and technology, fingerprint acquisition device appears in the every aspect of people's life.Such as: fingerprint attendance machine,
Fingerprint access control, mobile device etc..
In general, fingerprint acquisition device has obtained the finger print data of the secondary scanning, later, has passed through transmission after completing sampling
Bus exports data.But in some cases, directly transmission source data (raw data) will lead to system performance decline sum number
According to loss.
The common restricted occasion of some hardware is as follows:
1, there are when contradiction by transmitted data amount d, transmission rate rate and transmission time t.
The relationship of three is that d > rate × t can not be by one that is, in the case where given transmission time and transmission rate
Quantitative data all transfer out.When system requirements in a very short period of time, with a lower transmission rate, transmit one group
When larger amount of data, the case where just will appear performance degradation.At this point, being merely that can not solve this situation by hardware
, it must be adjusted from system level, such as increase transmission time or increase transmission rate, can influence whether system in this way and often
The work of other modules.
2, there are when contradiction by transmitted data amount d and memory capacity m.
The relationship of the two is d > m, that is, the data volume for transmitting out has been more than the memory capacity of system distribution.In low cost
System schema in, memory capacity be it is very valuable, hardware can not be in a suitable manner by data needed for whole systems
It deposits in limited memory, will lead to the loss of data.
Summary of the invention
The purpose of the present invention is to provide a kind of compression method of finger print data and devices, before not influencing system performance
It puts, source data is compressed, reaching reduces bandwidth, memory space and the requirement for improving safety.
In order to solve the above technical problems, embodiments of the present invention provide a kind of compression method of finger print data, include
Following steps:
Compression based on comentropy is carried out to source finger print data；
It is stored or is exported using compressed image data as collected finger print data to next processing unit.
Embodiments of the present invention additionally provide a kind of compression set of finger print data, include:
Compression module, for carrying out the compression based on comentropy to source finger print data；
Memory module, for being stored compressed image data as collected finger print data；Or
Output module, for exporting compressed image data to next processing unit.
Embodiment of the present invention in terms of existing technologies, by compressing to source finger print data data, reduces
Transmission bandwidth and memory space, and since compression & decompression method is grasped by sending and receiving end, there is certain encryption effect, mention
The high safety of transmission.
In addition, also including before the step of carrying out the compression based on comentropy to source finger print data:
It detects whether to receive the instruction for being used to indicate compression；If not receiving described instruction, directly by the source fingerprint
Data are stored or are exported as collected finger print data to next processing unit；If described instruction is received, then
Into the step of the compression carried out to source finger print data based on comentropy.
By detecting whether to receive the instruction for being used to indicate compression, can choose whether to compress source finger print data,
When the data volume of transmission is greater than the amount of storage of system distribution, the instruction for being used to indicate compression is can be generated in system, then starts
Source finger print data is carried out based on information entropy squeezing, then compressed finger print data is stored or exported to next processing
Unit；When the data volume of transmission is less than the amount of storage of system distribution, system can not generate the instruction for being used to indicate compression, then
It is directly stored or is exported to next processing unit using source finger print data as collected finger print data, so that fingerprint number
According to can be flexible and changeable in transmission process, the adaptability of product be improved
In addition, including following sub-step in the step of carrying out the compression based on comentropy to source finger print data:
Source finger print data is predicted, predicted image data is obtained；By the predicted image data and the source fingerprint
Data carry out asking poor, obtain difference image data；Feature decomposition is carried out to the difference image data；Selected characteristic value is maximum
Several characteristic values and the corresponding feature vector of each characteristic value are as compressed image data, wherein the characteristic value
Weight of the corresponding feature vector in comentropy.
Difference is sought by the prediction of the first step and second step, the calculation amount of feature decomposition in third step can be effectively reduced.
Due to there is many multiply-add operations in feature decomposition, if carrying out the realization of hardware based on this, register bit wide will be very big, leads
Cause the rising of cost.If carrying out feature decomposition with difference, due to difference often very little, the deposit of hardware can be thus substantially reduced
Device bit wide, and then cost is reduced, improve the adaptability of product.
In addition, carrying out feature decomposition to the difference image data includes following sub-step:
Using square M × M as the matrix G of difference image data described in unit cuttingdiff；The M obtained for each cutting
The submatrix of × M size, carry out singular value decomposition SVD, each submatrix obtain descending arrangement M characteristic value, and with the M
A one-to-one M group feature vector of characteristic value；
The selected characteristic is worth several maximum characteristic values and the corresponding feature vector of each characteristic value as compressed
Image data includes following sub-step:
The preceding k characteristic value in M characteristic value of each submatrix is taken, and corresponding with the preceding k characteristic value
Feature vector, several maximum characteristic values of characteristic value and the corresponding feature vector of each characteristic value as selection；Wherein, described
M, k is preset natural number, and the k is less than or equal to the M.
Pass through the matrix G to difference image datadiffProgress cutting, the available lesser submatrix of numerical value, and to more
Small submatrix carries out SVD (singular value decomposition), can substantially reduce the processing capacity requirement to system, improve computational efficiency.Separately
It can control the ratio of compression by the selection of k value outside, the smaller compression ratio of k value is bigger, otherwise the bigger compression ratio of k value is more
It is small, so that the process of compression is simply controllable.
In addition, described using square M × M as the matrix G of difference image data described in unit cuttingdiffThe step of it
Before, also include:
Judge the matrix G of the difference image datadiffSize whether be square the integral multiple of M × M；If it is determined that
As a result be it is no, then by the matrix GdiffPolishing is the integral multiple of M × M.
By matrix GdiffPolishing is that the integral multiple of M × M may insure to error image matrix GdiffObtain when cutting whole
Several submatrixs.
Detailed description of the invention
Fig. 1 is the hardware structural diagram of fingerprint data compression device in embodiment according to the present invention；
Fig. 2 is finger print data compression method flow chart according to first embodiment；
Fig. 3 be according to first embodiment in flow chart that image data is compressed；
Fig. 4 be according to first embodiment in when predicting image data the position of current pixel and neighborhood pixels show
It is intended to；
Fig. 5 be according to first embodiment in algorithm examples schematic diagram that image data is predicted；
Fig. 6 be according to first embodiment in matrix of differences GdiffQuantization schematic diagram；
Fig. 7 be according to first embodiment in difference image data carry out feature decomposition when different value of K under compression calculate
Method reconstruct image；
Fig. 8 is the finger print data compression method flow chart according to second embodiment；
Fig. 9 is the finger print data compression set block schematic illustration according to third embodiment.
Specific embodiment
To make the object, technical solutions and advantages of the present invention clearer, below in conjunction with attached drawing to each reality of the invention
The mode of applying is explained in detail.However, it will be understood by those skilled in the art that in each embodiment of the present invention,
In order to make the reader understand this application better, many technical details are proposed.But even if without these technical details and base
In the various changes and modifications of following embodiment, each claim of the application technical side claimed also may be implemented
Case.
Embodiments of the present invention propose a kind of hardware configuration for having finger print data compression set (compressor),
It converts source data to compressed data (compressed data) in some way, and be relatively easy in the realization of hardware and
Cost is more excellent.Its core technology includes:
1, hardware system level
As shown in Figure 1: after analog-digital converter (ADC), joined data compressing module, and can choose unlatching or
It closes.It when the module is opened, is then transmitted again after carrying out data processing using data compressing module, the data of output are after compressing
Data；When the module is closed, the data of output are source data.
2, based on the compression algorithm of feature decomposition
Common compression algorithm is quantified as leading with Fourier transformation or class Fourier transformation and combination, is the calculation based on frequency domain
Method.And technology of the invention is based on feature decomposition (Eigen decomposition), is the algorithm based on comentropy.But
It is that previous feature decomposition compression method does not account for the characteristic of image, i.e., image data is from space or in view of the time, all
There is very strong correlation.In embodiments of the present invention, the method for prediction is proposed, feature decomposition is carried out to difference, keeps its right
It is more easier in the realization of hardware and cost is lower.
3, under the premise of not influencing performance, using lossy compression
In view of specific application scenarios, handled without using source data come further, because after lossy compression again
Identical or very approximate data input can be provided by rebuilding data (decompressed data).
It is specifically described below for the embodiments of the present invention.
The first embodiment of the present invention is related to a kind of finger print data compression method, detailed process is as shown in Figure 2:
In step 201, fingerprint is acquired, the acquisition of fingerprint is such as carried out using fingerprint acquisition device.
In step 202, processing is amplified to collected finger print data, by enhanced processing, can makes to collect
Finger print data in weaker signal become strong, it is convenient that finger print data is further processed.
In step 203, the finger print data after enhanced processing is filtered, fingerprint number can be eliminated by filtering
Interference signal in improves the accuracy of finger print data processing.
In step 204, carrying out analog-to-digital conversion to the finger print data after filtering processing makes fingerprint number by analog-to-digital conversion
Digital quantity is converted to according to from analog quantity, finger print data is compressed and transmitted after convenient.
In step 205, the finger print data after analog-to-digital conversion is compressed, specifically, being compressed in present embodiment
The core concept of algorithm to source finger print data predict, seek difference after carry out Eigenvalues Decomposition, then select several characteristic values larger
Feature vector express source images (source finger print data), detailed process (i.e. the process that data compressing module in Fig. 1 is realized)
It is as shown in Figure 3:
In step 301, the finger print data after analog-to-digital conversion is predicted, obtains predicted image data.
Since image data is from space or in view of the time, there is very strong correlation, i.e., it is intraframe or interframe at some
The gray value of the gray value of position and close position is very close to, and relevant property, (in frame) as shown in Figure 4.Therefore, if known neighbour
The gray value of near position, so that it may predict the gray value of current location.Certainly, the gray value G of predictionpredIt is deposited with true value G
In difference Gdiff。
Prediction algorithm with formula 4-1 by being provided:
Without loss of generality, it will be assumed that a≤b, then the meaning of prediction algorithm can understand in this way: when the left side of x is deposited
At a vertical edge, then select pixel a as the predicted value of x；If there is a horizontal edge above x, picture is chosen
Predicted value of the plain b as x；If not detecting edge, with a planar prediction a+b-c as the predicted value of x.Its
In, edge refers to the point very big with the grey value difference of current pixel, and non-edge refers to close with the gray value of current pixel
Point.From the above, it is seen that this prediction has edge self-checking function, it is particularly suitable for fingerprint image, and realizes simple.
Assuming that picture size is w × h, each sampled point is N, such as N=12.We predict the number that a sub-picture needs
It is only (w+h-1) × N according to amount.
In step 302, it carries out the finger print data after predicted image data and analog-to-digital conversion to ask poor, obtains error image
Data.
The matrix G of difference image datadiffIt is provided by formula 4-2:
Gdiff=Gpred-G (4-2)
This prediction process is shown with the true fingerprint image of a width below, as shown in Figure 5: left figure is source images, in
Figure is forecast image, and right figure is error image.
In the matrix G to difference image datadiffDuring being quantified, it is assumed that each sampled point is N, such as N
=12, obtaining GdiffAfterwards, which has positive number also to have negative, and its absolute value is centainly less than 2N.In order to further refine
GdiffRange, the maximum value of its bit wide is set to round (2 × N/3+1) by analysis of experimental data by us, contain sign bit,
As shown in Figure 6: wherein left figure is source error image, and middle graph is error image after quantization, and right figure is the two error image.
In step 303, judge whether the size of the matrix of the difference image data is square the integral multiple of M × M,
The value range of middle M is 4 to 32, and preferably M can be taken as 16, may make and reaches preferable in the effect that calculation amount is approached with image
Balance.If it is judged that be it is yes, then be directly entered in step 304 and cutting carried out to the matrix of difference image data, otherwise
305 are entered step, by the matrix of difference image data with M × M polishing.
In step 304, by error image matrix GdiffCutting is carried out by unit of square M × M, obtains several height
Matrix.Specifically, several submatrix sizes obtained after cutting are M × M.
In step 305, by matrix GdiffPolishing is the integral multiple of M × M.If specifically, matrix GdiffSize not
It is the integral multiple of M × M, needs matrix of differences GdiffPolishing is the integral multiple of M × M, it can be ensured that matrix GdiffCarry out cutting
When obtain integer submatrix.
Within step 306, SVD (singular value decomposition) is carried out to a submatrix, obtains M characteristic value of descending arrangement,
And with the one-to-one M group feature vector of the M characteristic value.Specifically, the corresponding spy of the characteristic value obtained after SVD
Weight of the vector in comentropy is levied, being SVD to submatrix can be decomposed by QR (i.e. by matrix decomposition at an orthonomal matrix
The methods of Q and upper triangular matrix R) are completed, and the methods of being decomposed due to QR is universal method, therefore details are not described herein.In addition,
Descending arrangement is carried out to M obtained characteristic value, and with the one-to-one M group feature vector of the M characteristic value, can be convenient
Value is carried out to M characteristic value.
In step 307, the preceding k characteristic value in M characteristic value, and spy corresponding with preceding k characteristic value are taken
Vector is levied, it is stored, wherein k is less than or equal to M.The selection of k value is very crucial in this step, forces because it will be directly influenced
Close effect, while also influencing compression factor.Have when the value range of k is 2 to M under normal circumstances and relatively good approaches effect
Fruit.In present embodiment, k value is taken as 3, to reach the balance of preferable Approximation effect and compression ratio.Fig. 7 is to use this reality
The example images that the feature vector that the mode of applying is chosen is reconstructed, it is seen then that almost consistent with source images in k=3, explanation is forced
Nearly effect is fine.
By experiment, available k=ceil (M/6), wherein ceil function is toward positive value direction value, such as ceil
(2.3) 3 can be obtained.The step of from front it is found that work as M=16, when k=3, that is, take preceding 3 from the characteristic value that 16 descendings arrange
A and its corresponding 2 length is the feature vector of M, can reach preferable flat in the effect that calculation amount and image are approached
Weighing apparatus.
In step 308, judge whether that there is also the submatrixs for not carrying out SVD (singular value decomposition), specifically, if
Judging result be it is yes, illustrate to matrix GdiffCarry out the sub- square for not carrying out SVD in several submatrixs obtained after cutting also
Battle array then returns to step 306 and carries out SVD to next submatrix, finishes until all decomposing.If it is judged that being no explanation
All submatrixs all carry out SVD, then are directly entered step 309.
The submatrix of each M × M size obtained after cutting is all carried out SVD one by one, obtained by step 306 to 308
Take the preceding k characteristic value of each submatrix, and feature vector corresponding with preceding k characteristic value, the differential chart as selection
As several maximum characteristic values of characteristic value and the corresponding feature vector of each characteristic value in data.
In a step 309, by the preceding k characteristic value of submatrix each in all submatrixs, and distinguish with k characteristic value
Corresponding feature vector, i.e., by several maximum characteristic values of characteristic value in difference image data and the corresponding feature of each characteristic value
Vector, as compressed image data.
In step 206, compressed image data is stored as collected finger print data or is exported under
One processing unit.Such as it stores compressed image data as collected finger print data to self-purchased storage list
In member, alternatively, outputting data to processing by transfer bus using compressed image data as collected finger print data
In device or Installed System Memory.
It is not difficult to find that by carrying out prediction to the finger print data after analog-to-digital conversion and to predicted image data and analog-to-digital conversion
Finger print data afterwards carries out asking poor, it is possible to reduce the calculation amount of feature decomposition, due to difference very little, in the SVD carried out with difference
When decomposition, the register bit wide of hardware can be substantially reduced, cost is made more to have adaptability.It is compressed additionally by finger print data
Storage or transmission data volume can be reduced and compression ratio is controllable, since compression & decompression method is grasped by sending and receiving end, tool
There is certain encryption effect, improves the requirement of safety.
Second embodiment of the present invention is related to a kind of compression method of finger print data.Second embodiment is implemented first
It improves, thes improvement is that on the basis of mode: in the first embodiment, directly will be collected by compression algorithm
It is stored or is exported to next processing unit after finger print data compression.And in second embodiment of the invention, pass through inspection
The instruction for whether receiving and being used to indicate compression image is surveyed, chooses whether to compress collected finger print data.Detailed process
Such as Fig. 8 institute:
Step 801-804 is as the step 201-204 in first embodiment, and details are not described herein.
In step 805, it detects whether to receive the instruction for being used to indicate compression image, specifically, when the data of transmission
When amount is less than the amount of storage of system distribution, system can not be generated to the instruction for being used to indicate compression image, then enter step 806
Directly by the image data after analog-to-digital conversion, is stored or exported as collected finger print data and is single to next processing
Member.When the data volume of transmission is greater than the amount of storage of system distribution, the instruction for being used to indicate compression image is can be generated in system, then
807 are entered step to start to compress the image data after analog-to-digital conversion.
When passing through hardware specific implementation, a control signal is sent using an external register, control signal is used for
Whether the switch in control figure 1 is connected, whether compression to control the image data after analog-to-digital conversion.
Step 807-808 is as the step 205-206 in first embodiment, and details are not described herein.
Present embodiment passes through the instruction for detecting whether to receive and being used to indicate compression image, chooses whether to collected finger
Line data are compressed, and allow the transmission process of finger print data flexible and changeable, improve the adaptability of product.
The step of various methods divide above, be intended merely to describe it is clear, when realization can be merged into a step or
Certain steps are split, multiple steps are decomposed into, as long as comprising identical logical relation, all in the protection scope of this patent
It is interior；To adding inessential modification in algorithm or in process or introducing inessential design, but its algorithm is not changed
Core design with process is all in the protection scope of the patent.
Third embodiment of the invention is related to a kind of compression set of finger print data, includes:
Compression module, for carrying out the compression based on comentropy to source finger print data.
Memory module, for being stored compressed image data as collected finger print data；Or
Output module, for exporting compressed image data to next processing unit；And
Control module directly will for detecting whether receiving the instruction for being used to indicate compression, and when not receiving the instruction
Source finger print data is stored or is exported as collected finger print data to next processing unit；When receiving the instruction,
Triggering compression module carries out the compression based on comentropy to source finger print data.
It should be noted that also including analog-to-digital conversion module in present embodiment, for collected source finger print data
Analog-to-digital conversion is carried out, as shown in Figure 9.
In the present embodiment, compression module also includes:
Predict that submodule obtains predicted image data for predicting source finger print data；
Poor submodule is sought, for carrying out predicted image data and source finger print data to ask poor, obtains difference image data；
Feature decomposition submodule, for carrying out feature decomposition to difference image data；
Submodule is chosen, is worth several maximum characteristic values and the corresponding feature vector of each characteristic value for selected characteristic,
As compressed image data；Wherein, this feature value characterizes weight of the corresponding feature vector in comentropy.
It is not difficult to find that present embodiment is system embodiment corresponding with the compression algorithm in second embodiment, this
Embodiment can work in coordination implementation with the compression algorithm in second embodiment.It is mentioned in compression algorithm in second embodiment
The relevant technical details arrived are still effective in the present embodiment, and in order to reduce repetition, which is not described herein again.Correspondingly, this reality
It applies in the compression algorithm that the relevant technical details mentioned in mode are also applicable in second embodiment.
It is noted that each module involved in present embodiment is logic module, and in practical applications, one
A logic unit can be a physical unit, be also possible to a part of a physical unit, can also be with multiple physics lists
The combination of member is realized.In addition, in order to protrude innovative part of the invention, it will not be with solution institute of the present invention in present embodiment
The technical issues of proposition, the less close unit of relationship introduced, but this does not indicate that there is no other single in present embodiment
Member.
It will be understood by those skilled in the art that the respective embodiments described above are to realize specific embodiments of the present invention,
And in practical applications, can to it, various changes can be made in the form and details, without departing from the spirit and scope of the present invention.
Claims (7)
1. a kind of compression method of finger print data, which is characterized in that comprise the steps of:
Compression based on comentropy is carried out to source finger print data；
It is stored or is exported using compressed image data as collected finger print data to next processing unit；
Include following sub-step in the step of compression carried out to source finger print data based on comentropy:
Source finger print data is predicted, predicted image data is obtained；
It carries out the predicted image data and the source finger print data to ask poor, obtains difference image data；
Feature decomposition is carried out to the difference image data；
Selected characteristic is worth several maximum characteristic values and the corresponding feature vector of each characteristic value as compressed image data,
Wherein, weight of the corresponding feature vector of the characteristic value in comentropy；
Described includes following sub-step to difference image data progress feature decomposition:
Using square M × M as the matrix G of difference image data described in unit cuttingdiff；
For the submatrix for M × M size that each cutting obtains, singular value decomposition SVD is carried out, each submatrix obtains descending row
M characteristic value of column, and with the one-to-one M group feature vector of the M characteristic value；
The selected characteristic is worth several maximum characteristic values and the corresponding feature vector of each characteristic value as compressed image
Data include following sub-step:
Take the preceding k characteristic value in M characteristic value of each submatrix, and feature corresponding with the preceding k characteristic value
Vector, several maximum characteristic values of characteristic value and the corresponding feature vector of each characteristic value as selection；
Wherein, described M, k are preset natural number, and the k is less than or equal to the M.
2. the compression method of finger print data according to claim 1, which is characterized in that be based on to source finger print data
Before the step of compression of comentropy, also include:
It detects whether to receive the instruction for being used to indicate compression；
If not receiving described instruction, directly using the source finger print data as collected finger print data carry out storage or it is defeated
Out to next processing unit；
If receiving described instruction, enters back into and described the compression based on comentropy is carried out to source finger print data.
3. the compression method of finger print data according to claim 1, which is characterized in that it is described with square M × M be single
The matrix G of difference image data described in the cutting of positiondiffThe step of before, also include:
Judge the matrix G of the difference image datadiffSize whether be square the integral multiple of M × M；
If it is judged that be it is no, then by the matrix GdiffPolishing is the integral multiple of M × M.
4. the compression method of finger print data according to claim 1, which is characterized in that the value range of the M be 4 to
32。
5. the compression method of finger print data according to claim 1, which is characterized in that the value range of the k is 2 to institute
State M.
6. a kind of compression set of finger print data, characterized by comprising:
Compression module, for carrying out the compression based on comentropy to source finger print data；
Memory module, for being stored compressed image data as collected finger print data；Or
Output module: for exporting compressed image data to next processing unit；
The compression module includes:
Predict that submodule obtains predicted image data for predicting source finger print data；
Poor submodule is sought, for carrying out the predicted image data and the source finger print data to ask poor, obtains error image number
According to；
Feature decomposition submodule, for carrying out feature decomposition to the difference image data；
Submodule is chosen, is worth several maximum characteristic values and the corresponding feature vector of each characteristic value for selected characteristic, as
Compressed image data；Wherein, weight of the corresponding feature vector of the characteristic value in comentropy；
The feature decomposition submodule, is specifically used for:
Using square M × M as the matrix G of difference image data described in unit cuttingdiff；
For the submatrix for M × M size that each cutting obtains, singular value decomposition SVD is carried out, each submatrix obtains descending row
M characteristic value of column, and with the one-to-one M group feature vector of the M characteristic value；
The selection submodule, is specifically used for:
Take the preceding k characteristic value in M characteristic value of each submatrix, and feature corresponding with the preceding k characteristic value
Vector, several maximum characteristic values of characteristic value and the corresponding feature vector of each characteristic value as selection；
Wherein, described M, k are preset natural number, and the k is less than or equal to the M.
7. the compression set of finger print data according to claim 6, which is characterized in that also include:
Control module, for detecting whether the instruction for being used to indicate compression is received, and when not receiving described instruction, directly by institute
Source finger print data is stated as collected finger print data and is stored or is exported to next processing unit；Receiving described instruction
When, it triggers the compression module and the compression based on comentropy is carried out to the source finger print data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610011163.9A CN106954002B (en) | 2016-01-07 | 2016-01-07 | A kind of compression method and device of finger print data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610011163.9A CN106954002B (en) | 2016-01-07 | 2016-01-07 | A kind of compression method and device of finger print data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106954002A CN106954002A (en) | 2017-07-14 |
| CN106954002B true CN106954002B (en) | 2019-03-26 |
Family
ID=59465659
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610011163.9A Active CN106954002B (en) | 2016-01-07 | 2016-01-07 | A kind of compression method and device of finger print data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106954002B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110113708B (en) * | 2018-04-18 | 2020-12-01 | 爱动超越人工智能科技（北京）有限责任公司 | Positioning method and device based on Wi-Fi position fingerprint |
| EP3852015A4 (en) * | 2018-09-30 | 2021-09-01 | Huawei Technologies Co., Ltd. | Operational accelerator and compression method |
| TWI759818B (en) * | 2020-08-11 | 2022-04-01 | 國立高雄科技大學 | Method and system for detecting singular points in fingerprint images with entropy-based clustering algorithmic processing |
| CN112199049B (en) * | 2020-10-22 | 2023-10-20 | Tcl通讯（宁波）有限公司 | Fingerprint storage method, fingerprint storage device and terminal |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2007107050A1 (en) * | 2006-03-23 | 2007-09-27 | Zksoftware Beijing Inc. | Fingerprint identification method and system |
| CN103942566A (en) * | 2014-04-15 | 2014-07-23 | 清华大学 | Compressing method and system of fingerprint descriptors |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9070002B2 (en) * | 2011-10-18 | 2015-06-30 | King Fahd University Of Petroleum And Minerals | Method of performing fingerprint matching |
-
2016
- 2016-01-07 CN CN201610011163.9A patent/CN106954002B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2007107050A1 (en) * | 2006-03-23 | 2007-09-27 | Zksoftware Beijing Inc. | Fingerprint identification method and system |
| CN103942566A (en) * | 2014-04-15 | 2014-07-23 | 清华大学 | Compressing method and system of fingerprint descriptors |
Non-Patent Citations (5)
| Title |
|---|
| 图像通信中SVD压缩算法研究;张立新;《江苏科技信息》;20110831;前言部分-第5节、图2 |
| 基于块方向信息的指纹图像无损压缩编码;唐良瑞，蔡安妮，孙景鳌;《计算机工程》;20021130;前言部分-第4节误差图像的熵编码、附图1 |
| 基于块方向信息的指纹图像无损压缩编码;唐良瑞，蔡安妮，孙景鳌;《计算机工程》;20021130;同上 |
| 基于奇异值分解(SVD)的图像压缩;胡乡峰，卫金茂;《东北师大学报》;20060930;全文 |
| 指纹图像压缩方法的研究;马义德,段磊,韩明秋,刘悦;《甘肃科学学报》;20050630;全文 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106954002A (en) | 2017-07-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106954002B (en) | A kind of compression method and device of finger print data | |
| CN104838653B (en) | Lossless image compression using differential transmission | |
| EP3211894B1 (en) | Method and apparatus for performing graph-based prediction by using optimization function | |
| US8805106B2 (en) | System and method for compressing and decompressing images and video | |
| Khashman et al. | Image compression using neural networks and Haar wavelet | |
| Li et al. | GPU acceleration of clustered DPCM for lossless compression of hyperspectral images | |
| Lee et al. | Visually weighted compressive sensing: Measurement and reconstruction | |
| Zhao et al. | Image compressive-sensing recovery using structured laplacian sparsity in DCT domain and multi-hypothesis prediction | |
| JPH0974543A (en) | Image signal conversion device and method | |
| Korde et al. | Wavelet based medical image compression for telemedicine application | |
| Monika et al. | Adaptive block compressed sensing-a technological analysis and survey on challenges, innovation directions and applications | |
| CN106797447A (en) | Color index figure compression set and method | |
| Deshlahra | Analysis of Image Compression Methods Based On Transform and Fractal Coding | |
| Ramesh et al. | Analysis of lossy hyperspectral image compression techniques | |
| Li et al. | Optimization of integer wavelet transforms based on difference correlation structures | |
| CN108933942A (en) | A kind of filtering method compressing video and the filter for compression video | |
| Naqvi et al. | Sparse representation of image and video using easy path wavelet transform | |
| Sarawadekar et al. | Area efficient, high-speed VLSI design for EBCOT block coder in JPEG 2000 | |
| Fowler et al. | Classification performance of random-projection-based dimensionality reduction of hyperspectral imagery | |
| CN112616052B (en) | Method for reconstructing video compression signal | |
| Agarwal et al. | Image Compression Techniques Comparative Analysis using SVD-WDR and SVD-WDR with Principal Component Analysis | |
| Hua et al. | Wavelet-based coding of time-varying vector fields of ocean-surface winds | |
| WO2020207451A1 (en) | H.265 encoding method and apparatus | |
| EP2254338B1 (en) | Coding and decoding methods and devices, computer program and information carrier enabling the implementation of such methods | |
| Pawar et al. | A review: Approximate image and video compression |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |