The computational methods of peak frequency envelope curve and device
Technical field
The present invention relates to medical ultrasonic technical field, particularly relate to computational methods and the device of peak frequency envelope curve.
Background technology
In medical ultrasonic technical field, utilize Doppler effect to obtain spectrogram thus measure the method for blood flow velocity and obtain in recent years and develop rapidly.
Ultrasonic Doppler technique is measured in the process of blood flow rate, and continuous ultrasound doppler system sends ultrasound wave in tissue through probe, and receives the scatter echo that hemocyte causes.Due to the mobility of hemocyte, have certain frequency shift (FS) between its scatter echo and transmitted wave, and the movement velocity of hemocyte being proportional to this frequency shift (FS), is be Doppler effect.
Measure corresponding frequency deviation and can estimate endovascular blood circumstance.Measure in the process of blood flow rate at pulse Doppler, measuring system sends sound wave pulse according to certain time interval, and echo-signal is received between two subpulses, measure the phase contrast of echo and certain reference signal, this phase contrast is the same with the frequency deviation in continuous-wave doppler, and rate of change is proportional to the flowing velocity of blood flow.Process is continuous is clinically consistent with the method for pulse Doppler.
Echo-signal, after the process such as orthogonal modulation, filtering, becomes the audio Doppler centered by 0 frequency, about a few KHz of its bandwidth.The flow velocity of blood in human vas is unfixed, and diverse location has different speed.General angiocentric flow velocity is comparatively large, and the local flow velocity of close blood vessel wall is close to 0.Therefore Doppler signal frequency spectrum has certain width, wherein peak frequency is proportional to endovascular peak veloity,PV, mid frequency (prominent frequency) represents the flow tendency of vessel inner blood entirety, and the erythrocyte with this flow velocity accounts for maximum ratio in blood.By the impact of heart contraction and diastole, systems axiol-ogy to Doppler signal frequency spectrum also change thereupon.The different frequency composition of synchronization is modulated into gray scale by watt level, obtain the spectrum line in this moment, if repeatedly intercept Doppler signal by certain hour interval to carry out frequency spectrum calculating, and temporally order shows spectrum line successively, just obtain Doppler signal sound spectrogram.
Peak frequency envelope curve, extracts and obtains from sound spectrogram.Can obtain by connecting not the peak frequency of Doppler signal frequency spectrum in the same time, this profile amplitude is directly proportional to Ink vessel transfusing blood flow Peak Flow Rate.Described peak frequency is estimation average frequency and the basis calculating other Doppler parameter, therefore estimates that peak frequency is in important significance in clinical application exactly.
The method of initial estimation peak frequency envelope curve is: determine by operator is manual in freeze frame, then calculate other parameters corresponding.Clearly, poor repeatability, precision is low, also cannot do real-time estimation for this method shortcoming.
At present, the method extracting peak frequency envelope curve mainly contains two classes: a class is the processing method based on image, as rim detection, and dimension modulation Z axis method etc., the shortcoming of these class methods is that amount of calculation is too large, is difficult to application under real-time conditions; Another kind of is utilize power spectral density integral curve, carries out peak frequency estimation, accounts for dominant position in actual applications, as percentage method (PM), crosses threshold method (TCM), geometric method (GM) etc.
Percentage method is setting peak frequency is a percentage ratio of signal plus noise power summation, calculate relatively simple, but it is responsive to signal to noise ratio, is difficult to determine a stable percentage ratio, and optimum percentage ratio is in the same time not inconsistent yet.Cross threshold method by setting threshold value, to find in a continuous n Frequency point point that at least m power spectral value exceedes this threshold value, the threshold value setting of this method is very crucial, and threshold value then Maximum frequency curve bigger than normal is less than normal, and threshold value is too small, easily noise detected.Geometric method carrys out matching integrated power spectrum with three straight lines, and it does not need selected threshold, but it is larger to the dependency of spectrum peak, and the peak utilizing FFT to try to achieve has very large randomness, have impact on the estimated accuracy of peak frequency.
On the basis of these three kinds of methods, some innovatory algorithm are derived, threshold method (MTCM) is crossed as what improve, mixing method (HM), the geometric method (MGM) improved, Adaptive Thresholding (ATM) and the percentage method (MPM) etc. improved, improve the effect of algorithm to a certain extent, but do not have large progress.
The shortcoming of prior art: the accuracy of existing percentage ratio algorithm is subject to the restriction of signal noise ratio (snr) of image, and the accuracy of percentage ratio algorithm is not high.
Summary of the invention
Technical problem to be solved by this invention is: how the signal to noise ratio of sound spectrogram picture to be associated with the percentage ratio value of percentage ratio algorithm, improves the accuracy of percentage ratio algorithm.
In an embodiment of the present invention, propose the computational methods of peak frequency envelope curve, comprising:
S1, the single spectral line calculating sound spectrogram from baseline up have an energy accumulation and sum;
S2, judges whether sum is greater than first threshold; Get percent value and equal β, otherwise, get percent value and equal α, wherein β > α;
S3, calculate single spectral line from baseline up to current point energy accumulation and sum1, judge whether sum1 is greater than the product of the percent value described in sum and S2, is, stopping calculating sum1, getting current point is the point that on this spectrum line, energy accumulation is maximum; Otherwise, calculate sum1, until sum1 is greater than the product of the percent value described in sum and S2;
S4, to calculating by root spectrum line on sound spectrogram, obtains the point that on every root spectrum line, energy accumulation is maximum;
S5, connects the point that on all spectrum lines of sound spectrogram, energy accumulation is maximum and forms peak frequency envelope curve.
In addition, further embodiment of this invention additionally provides the accountant of peak frequency envelope curve, comprising:
Line gross energy computing unit, for the single spectral line that calculates sound spectrogram from baseline up have an energy accumulation and sum;
Percent value determining unit, for judging whether described sum is greater than first threshold; Get percent value and equal β, otherwise, get percent value and equal α, wherein β > α;
The computing unit of the point that energy accumulation is maximum on line, for calculate single spectral line from baseline up to current point energy accumulation and sum1, judge whether sum1 is greater than the product of sum and described percent value, is, stop calculating sum1, then current point is the point that on this spectrum line, energy accumulation is maximum; Otherwise, calculate sum1, until sum1 is greater than the product of sum and described percent value;
The computing unit of the point that energy accumulation is maximum on image, for calculating by root spectrum line on sound spectrogram, obtains the point that on every root spectrum line, energy accumulation is maximum;
Peak frequency envelope curve output unit, for being coupled together by point maximum for energy accumulation on all spectrum lines, forms peak frequency envelope curve.
As can be seen from the above technical solutions, the embodiment of the present invention has the following advantages: the signal to noise ratio of sound spectrogram picture associated with the percentage ratio value of percentage ratio algorithm, improve the accuracy of percentage ratio algorithm.
Accompanying drawing explanation
Fig. 1 is the flow chart of the computational methods of embodiment one peak frequency envelope curve;
Fig. 2 is the flow chart of the computational methods of preferred embodiment;
Fig. 3 is the accountant block diagram of embodiment two peak frequency envelope curve;
Fig. 4 is the accountant block diagram of preferred embodiment;
Fig. 5 is the implementation result figure of preferred embodiment.
Detailed description of the invention
Below in conjunction with the Figure of description in the present invention, be clearly and completely described the technical scheme in invention, obviously, described embodiment is only the present invention's part embodiment, instead of whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art, not making the every other embodiment obtained under creative work prerequisite, belong to the scope of protection of the invention.
embodiment one
In the present embodiment, see Fig. 1, the computational methods of peak frequency envelope curve, comprising:
S1, the single spectral line calculating sound spectrogram from baseline up have an energy accumulation and sum;
Sound spectrogram is made up of the many spectrum lines arranged according to time sequencing, all spectrum line upper frequencies be 0 point form baseline, every root spectrum line represents different frequency values from baseline difference up, has different gray scales simultaneously, represents different power.By the impact of heart contraction and diastole, the sound spectrogram of the Doppler signal that systems axiol-ogy arrives also thereupon in rhythmic change, as shown in Figure 5.
For calculating sound spectrogram on single spectral line from baseline up have a computational methods for energy accumulation sum, can obtain from prior art.
S2, judges whether sum is greater than first threshold; Get percent value and equal β, otherwise, get percent value and equal α, wherein β > α;
Described first threshold is an empirical value, and experimentally result is chosen, and preferably, described first threshold is 4000.
S3, calculate single spectral line from baseline up to current point energy accumulation and sum1, judge whether sum1 is greater than the product of the percent value described in sum and S2, is, stopping calculating sum1, getting current point is the point that on this line, energy accumulation is maximum; Otherwise, calculate sum1, until sum1 is greater than the product of the percent value described in sum and S2;
When sum1 is greater than the product of the percent value described in sum and S2, getting current point is the point that on this line, energy accumulation is maximum.
S4, to calculating by root spectrum line on sound spectrogram, obtains the point that on every root spectrum line, energy accumulation is maximum;
For the every root spectrum line on sound spectrogram, according to the method described in S1 to S3, calculate the point that energy accumulation on this spectrum line is maximum.
S5, connects the point that on all spectrum lines of sound spectrogram, energy accumulation is maximum, forms peak frequency envelope curve.See Fig. 2, preferably, between described S1 and S2, also S12 is comprised:
Judge whether the ratio of energy in integral image region and T and theoretical ceiling capacity and M is less than the first empirical value;
Choose first group of percent value;
Otherwise, judge whether the ratio of energy in grass energy and P and integral image region and T is less than the second empirical value; Choose first group of percent value; Otherwise, choose second group of percent value;
α and β of described first group of percent value is greater than α and β of second group of percent value respectively.
Described first empirical value experimentally result is chosen, and preferred first empirical value is 0.07.
Energy in described grass energy and P, integral image region and T, theoretical ceiling capacity and M, the explanation that its implication sees below.The feature of sound spectrogram is, under normal circumstances, change between adjacent spectral line is slower, and single spectral line is due to the impact by speckle noise, adding up the signal to noise ratio fluctuation obtained can be larger, if the image that the spectrum line recorded in a period of time is formed is considered as a whole, greatly can reduce the fluctuation of signal to noise ratio, improve system accuracies; In addition, the situation slower due to blood flow rate is in the majority, and most signal concentrates near baseline, and when paradoxical expansion blood flow rate is accelerated, image just occurs peak signal away from b extent.Grass is defined as 1/5 part on image top, namely, the part on each single spectral line top 1/5 all falls into grass region, in this part, generally the overwhelming majority all be full of by noise, although or have signal, the ratio of signal is lower, the energy in this subregion and be grass energy and P.Energy in computed image overall region and, be T.Theoretical ceiling capacity and M, when referring to that on hypothesis image, all points all have ceiling capacity, institute have an energy accumulation with.The energy in each region described in calculating and, each point energy in each corresponding for described region single spectral line is added.The ratio of described P and T represents the contrast situation of noise and signal, and ratio is less, illustrates that the intensity difference of signal and noise is larger, more easily distinguishes.The ratio of T and M, the signal intensity of representative image entirety, ratio is larger, then represent that signal is stronger, energy is higher, and ratio is little, illustrates that the intensity of echo is low.
When the ratio of the energy in integral image region and T and theoretical ceiling capacity and M is less than the first empirical value 0.07, or when but although this ratio ratio of being more than or equal to energy in the first empirical value 0.07 grass energy and P and integral image region and T is less than the second empirical value 0.018, it is comparatively clear that the signal band of key diagram picture and grass contrast, choose first group of percent value α be 0.94 and β be 0.97;
When the ratio of the energy in integral image region and T and theoretical ceiling capacity and M is greater than the first empirical value 0.07 and the ratio of energy in grass energy and P and integral image region and T is greater than the second empirical value 0.018, illustrate that image intensity signal is more weak, and signal intensity and noise intensity contrast not quite, choose second group of percent value α be 0.93 and β be 0.95.
In summary, these two groups of percent value, according to different image conditions choice for uses.
In the present embodiment, as T/M<0.07, choose first group of percent value α=0.94, β=0.97; When T/M >=0.07, and during P/T<0.018, choose first group of percent value α=0.94, β=0.97; And when T/M >=0.07, and during P/T >=0.018, choose second group of percent value α=0.93, β=0.95.
See Fig. 2, preferably, between described S1 and S12, also comprise
S11, judges whether sum is greater than Second Threshold, is, performs S12 step; Otherwise, peak frequency envelope curve is set to baseline.
Described Second Threshold is also an empirical value, and experimentally result is selected, and in the present embodiment, described Second Threshold is 400.If sum is less than Second Threshold, then illustrate that image signal energy is too small, peak frequency envelope curve is set to baseline.
To sum up, according to the signal to noise ratio of sound spectrogram, by organizing the selection of percent value, the accuracy of detection improving percentage ratio algorithm can be improved more; The present invention can obtain peak frequency envelope curve more accurately.
embodiment two
See Fig. 3, it should be noted that, the accountant of peak frequency envelope curve, comprising:
Line gross energy computing unit 21, for the single spectral line that calculates sound spectrogram from baseline up have an energy accumulation and sum;
Percent value determining unit 22, for judging whether described sum is greater than first threshold; Get percent value and equal β, otherwise, get percent value and equal α, wherein β > α;
The computing unit 23 of the point that energy accumulation is maximum on line, for calculate single spectral line from baseline up to current point energy accumulation and sum1, judge whether sum1 is greater than the product of sum and described percent value, is, stop calculating sum1, getting current point is the point that on this line, energy accumulation is maximum; Otherwise, calculate sum1, until sum1 is greater than the product of sum and described percent value;
The computing unit 24 of the point that energy accumulation is maximum on image, for calculating by root spectrum line on sound spectrogram, obtains the point that on every root spectrum line, energy accumulation is maximum;
Peak frequency envelope curve output unit 25, for being coupled together by point maximum for energy accumulation on all spectrum lines, forms peak frequency envelope curve.
See Fig. 4, it should be noted that, between described line gross energy computing unit 21 and percent value determining unit 22, also comprise: group chooses unit 212, for judging whether the ratio of energy in integral image region and T and theoretical ceiling capacity and M is less than the first empirical value; Choose first group of percent value; Otherwise, then judge whether the ratio of energy in grass energy and P and integral image region and T is less than the second empirical value, is choose first group of percent value; Otherwise, choose second group of percent value; α and β of described first group of percent value is greater than α and β of second group of percent value respectively.
See Fig. 4, it should be noted that, online gross energy computing unit 21 and group choose unit 212
Between, also comprise:
Low-yield judging unit 211, for judging whether sum is greater than Second Threshold, is, performing group and choosing unit; Otherwise, peak frequency envelope curve is set to baseline.
For device disclosed in embodiment, because it corresponds to the method disclosed in Example, so relevant part illustrates see method part.
The embodiment of the present invention has the following advantages:
The signal to noise ratio of sound spectrogram picture associates with the percentage ratio value of percentage ratio algorithm by the present invention, improves the accuracy of percentage ratio algorithm.See Fig. 5, the carotid artery Doppler frequency spectrum signal utilizing above-mentioned technology paired pulses Doppler ultrasound system to detect carries out envelope extraction experiment, and its Maximum frequency curve, as shown in White curves in figure, has well followed the tracks of the spectrum envelope of sound spectrogram.
Above the computational methods of peak frequency envelope curve provided by the present invention and device are described in detail, for one of ordinary skill in the art, according to the thought of the embodiment of the present invention, all will change in specific embodiments and applications, in sum, this description should not be construed as limitation of the present invention.