Elastography-based arterial vascular plaque characteristic test method and apparatus and storage medium

Abstract

The present invention provides an elastography-based arterial vascular plaque characteristic test method and apparatus and a storage medium. The method comprises: acquiring a tissue region of interest comprising plaque features in an arterial blood vessel; according to a tissue movement frequency, transmitting and receiving ultrasonic signals in the tissue region of interest; acquiring, from the ultrasonic signals, coarse and fine strain force image frames by means of calculation; fusing and matching the coarse and fine strain force image frames to acquire a precise strain force image frame, and acquiring a precise strain force image frame; counting the number of times each pixel in the precise strain force image frame has a strain force value for strain force, and performing image processing on counting results to obtain a feature image; and performing feature analysis on the feature image to obtain a feature analysis result, and performing output and display. The present invention improves the precision of plaque testing in arterial blood vessels on the basis of minimizing the amount of data, calculates, on the basis of strain information of the plaque, various characteristic indicator values ​​of the plaque, conveniently and quickly transmits same to a back-end for processing, and displays same on a display for user diagnosis.

Description

Arterial plaque characteristic detection method, device and storage medium based on elastic imaging Technical Field

[0001] The present invention relates to the technical field of ultrasonic medical equipment, and in particular to a method, device and storable medium for detecting arterial vascular plaque characteristics based on elastic imaging. Background Art

[0002] Carotid artery plaques, a manifestation of carotid atherosclerosis, are common at the bifurcation of the common carotid artery and are closely associated with ischemic stroke in the elderly. Medical color Doppler ultrasound (CDU) is the preferred noninvasive method for carotid artery examination and is widely used for screening and follow-up of carotid atherosclerosis. It not only demonstrates the location and size of plaques, as well as the location and severity of luminal stenosis, but also allows for hemodynamic assessment and morphological evaluation of plaques. Based on their morphology and echogenicity, plaques can be categorized as hypoechoic, lipid-rich soft plaques; moderately echogenic, collagen-rich, flat fibrous plaques; hyperechoic, calcified, hard plaques with acoustic shadowing; and mixed ulcerative plaques with varying echogenicity. Existing ultrasound methods for plaque assessment using CDU include IMT (intima-media thickness) measurement, which can quantitatively determine the presence of carotid atherosclerosis. However, assessment of carotid plaque characteristics still relies on subjective predictions based on hemodynamics, grayscale images, or morphology. In summary, in medical color Doppler ultrasound systems, the diagnosis of intra-arterial plaques is generally based on the echo strength, shape and size, assisted by hemodynamic characteristics for a subjective comprehensive diagnosis.

[0003] Therefore, a solution is urgently needed.

[0004] Summary of the Invention

[0005] One of the purposes of the present invention is to provide an arterial plaque characteristic detection method based on elastic imaging, which can be conveniently and quickly deployed in a medical color ultrasound equipment system.

[0006] The method for detecting arterial plaque characteristics based on elastic imaging provided by an embodiment of the present invention is applied to medical color Doppler ultrasound equipment, including:

[0007] Acquiring a tissue region of interest containing plaque features within an arterial vessel;

[0008] Transmit and receive ultrasonic signals in the tissue area of ​​interest according to the tissue movement frequency;

[0009] Obtaining coarse and fine strain force image frames by calculating ultrasonic signals;

[0010] fusing and matching the coarse and fine strain image frames to obtain precise strain image frames;

[0011] Counting the number of times each pixel in the precise strain image frame has a strain value, and performing image processing on the statistical results to obtain a feature image;

[0012] Perform feature analysis on the feature image, obtain feature analysis results, and output them for display.

[0013] Preferably, when obtaining a tissue region of interest containing plaque features within an arterial vessel, the ultrasound software is controlled to automatically locate the arterial plaque region of the current image as the tissue region of interest, or the user can manually select the region of interest.

[0014] Preferably, the scanning sampling frequency of transmitting and receiving ultrasonic signals is more than twice the frequency of tissue movement, and the scanning sampling frequency of transmitting and receiving is adjustable by the user.

[0015] Preferably, the ultrasonic signal is received and two or more different methods are used to calculate and obtain coarse and fine strain image frames, wherein the strain image method calculated by the high-precision slow method is a fine image, and the strain image method calculated by the low-precision fast method is a coarse image.

[0016] Preferably, the coarse and fine strain image frames are fused and matched, and the fusion and matching includes matching of positions of each pixel point between the coarse and fine image frames and superposition of strain forces.

[0017] Preferably, the temporal variation pattern of any pixel in the precise strain image frame can be obtained. The precise strain image frame is a temporally continuous frame. The system can automatically or manually select quantitative measurement points on the continuous frame. The ultrasound software can display and draw the numerical value, curve, or chart of the quantitative measurement point changing over time.

[0018] Preferably, the number of times a certain pixel has a certain strain value is counted for the precise strain image containing the plaque, and the statistical process can be performed in a frozen state of the image or in a real-time state.

[0019] Preferably, the strain corresponding to a certain pixel includes a pixel point collective strain consisting of one or two or more pixel points.

[0020] Preferably, data processing is performed on the statistical results to form a characteristic image, wherein the data processing is a process of normalizing the statistical times or recalculating the probability distribution.

[0021] Preferably, feature analysis of the feature image can obtain various types of mathematical or image feature indicators. The feature analysis is based on the feature image after normalization or re-probability distribution. The mathematical or image feature indicators include but are not limited to brightness, energy, contrast, uniformity, non-uniformity, homogeneity, non-homogeneity, correlation, non-correlation and other Chinese or English abbreviations.

[0022] Preferably, the final strain image and characteristic index are displayed comprehensively and output to a display for user diagnosis, wherein the characteristic index can be displayed dynamically or statically at any position on the display in the form of numerical values, charts, curves but not limited to such forms.

[0023] The arterial plaque characteristic detection device based on elastic imaging provided by the present invention is applied to medical color Doppler ultrasound equipment, comprising:

[0024] Select the tissue region of interest module to automatically obtain the manually obtained carotid artery plaque area;

[0025] Transmitter and receiver modules are used to obtain ultrasonic signals of tissue deformation;

[0026] A coarse and fine strain image frame calculation module is used to calculate multiple frames of coarse and fine strain images;

[0027] A precise strain image frame calculation module, used to calculate precise strain image frames;

[0028] A feature image calculation module is used to calculate the number of times a certain pixel has a certain strain value in response to the strain;

[0029] Parameter acquisition module, used to extract mathematical or image feature indicators from feature images;

[0030] The display module is used to display the final strain image and characteristic index values.

[0031] The present invention provides a computer-storable medium, which is used to execute the above-mentioned method for detecting arterial vascular plaque characteristics based on elastic imaging.

[0032] This application has achieved the following beneficial effects:

[0033] This patent enables quantitative measurement and qualitative analysis of plaque within arterial vessels. This patent utilizes elastic imaging technology from medical ultrasound equipment to improve the accuracy of plaque detection within arterial vessels while minimizing data volume. The patent also calculates various plaque characteristic indicators based on plaque strain information, conveniently and quickly transmitting this information to the backend for processing and display on a monitor for user diagnosis.

[0034] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present invention. The purposes and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings.

[0035] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0037] FIG1 is a schematic diagram of a method for detecting arterial plaque characteristics based on elastography according to an embodiment of the present invention;

[0038] FIG2 is a schematic diagram of an arterial plaque characteristic detection device based on elastography according to an embodiment of the present invention;

[0039] FIG3 is a schematic diagram of calculating coarse and fine strain image frames according to an embodiment of the present invention;

[0040] FIG4 is a schematic diagram of the definition of coarse and fine strain force image frames according to an embodiment of the present invention;

[0041] FIG5 is a schematic diagram of a method for fusing a certain plaque in a coarse and fine strain image according to an embodiment of the present invention;

[0042] FIG6 is a schematic diagram of real-time display of parameter indicators in an embodiment of the present invention;

[0043] FIG. 7 is a schematic diagram of the final display in an embodiment of the present invention. DETAILED DESCRIPTION

[0044] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0045] An embodiment of the present invention provides an arterial plaque characteristic detection method based on elastography, which is applied to medical color Doppler ultrasound equipment, including:

[0046] Acquiring a tissue region of interest containing plaque features within an arterial vessel;

[0047] Transmit and receive ultrasonic signals in the tissue area of ​​interest according to the tissue movement frequency;

[0048] Obtaining coarse and fine strain force image frames by calculating ultrasonic signals;

[0049] fusing and matching the coarse and fine strain image frames to obtain precise strain image frames;

[0050] Counting the number of times each pixel in the precise strain image frame has a strain value, and performing image processing on the statistical results to obtain a feature image;

[0051] Perform feature analysis on the feature image, obtain feature analysis results, and output them for display.

[0052] When acquiring a tissue region of interest containing plaque features within an arterial vessel, the ultrasound software is controlled to automatically locate the arterial plaque region in the current image as the tissue region of interest. The user can also manually select the region of interest.

[0053] The scanning sampling frequency of transmitting and receiving ultrasonic signals is more than twice the frequency of tissue movement, and the scanning sampling frequency of transmitting and receiving is adjustable by the user.

[0054] The ultrasonic signal can be received by using two or more different methods to calculate and obtain coarse and fine strain image frames, wherein the strain image method calculated by the high-precision slow method is the fine image, and the strain image method calculated by the low-precision fast method is the coarse image.

[0055] The coarse and fine strain image frames are fused and matched, and the fusion matching includes the matching of each pixel position between the coarse and fine image frames and the superposition of strain.

[0056] The temporal variation patterns of any pixel in the precise strain image frame can be obtained. The precise strain image frame is a continuous frame in time. The system can automatically or manually select quantitative measurement points on the continuous frame. The ultrasound software can display and draw the numerical values, curves or charts of the quantitative measurement points changing over time.

[0057] The number of times a certain pixel has a certain strain value for the precise strain image containing the plaque is counted, and the statistical process can be calculated in a frozen state or in a real-time state.

[0058] The strain corresponding to a certain pixel includes the strain of a pixel set consisting of one or two or more pixels.

[0059] The statistical results are processed to form a characteristic image, wherein the data processing is a process of normalizing the statistical times or recalculating the probability distribution.

[0060] Feature analysis of feature images can obtain various types of mathematical or image feature indicators. Feature analysis is based on the feature images after normalization or re-probability distribution. Mathematical or image feature indicators include but are not limited to brightness, energy, contrast, uniformity, non-uniformity, homogeneity, non-homogeneity, correlation, non-correlation and other Chinese or English abbreviations.

[0061] The final strain image and characteristic index are comprehensively displayed and output to a display for user diagnosis, wherein the characteristic index can be displayed dynamically or statically at any position on the display in the form of numerical values, charts, curves but not limited to such forms.

[0062] First, let’s explain the proper nouns involved, as follows:

[0063] Beamforming: Combining multiple ultrasonic signals into one signal is called beamforming;

[0064] Delay parameter: When multiple ultrasonic waves arrive at the target point, there is a sequence of arrival. This time difference is the delay parameter.

[0065] Dynamic reception: In medical color Doppler ultrasound equipment, the direction of ultrasound transmission is artificially defined as the axial direction. Ultrasound is recovered at short intervals along the axial direction. This reception method is called dynamic reception.

[0066] Focusing: Multiple ultrasonic signals arrive at a certain axial depth at the same time, which is called focusing;

[0067] Beam: The ultrasonic signal after beam synthesis is called a beam, which can also be called a beam line or line;

[0068] RF data: the final data after beamforming.

[0069] Specifically, the implementation case includes the following steps:

[0070] As shown in step S11 of Figure 1, the input data of the tissue region of interest of the intra-arterial plaque feature is defined as Pre-Frame(t-1, i, j) or Post-Frame(t, i, j) to represent the previous frame or the next frame, as shown in S30 or S31 of Figure 3, where t represents the frame number at time t; i represents the horizontal pixel label; and j represents the vertical pixel label. The left and right boundaries of the region of interest are [-L:1:L], i.e., with a step size of 1, the left boundary position is -L, and the right boundary position is L; the upper and lower boundaries are defined as [-H:1:H], i.e., with a step size of 1, the upper boundary is defined as -H, and the lower boundary is defined as H.

[0071] As shown in step S12 of Figure 1, the front-end transmitting and receiving modules perform direct scanning in a sequential manner. After scanning the grayscale image, they scan the tissue within the regions of interest (ROIs) [-L:1:L] and [-H:1:H], repeating this cycle. The repetition frequency of the transmitted and received ultrasonic signals is adjustable, set to the FRF, which is generally at least twice the tissue deformation frame rate.

[0072] As shown in step S13 of Figure 1, the strain image of the tissue can be calculated by the displacement difference between the previous and next two frames. As shown in Figure 3, the plaque S32 moves to the position S33. The position of the previous frame S30 is S35, represented by S0. S0 can be regarded as the center of mass of the plaque. The position in the next frame S31 is S2. When the plaque is deformed, the center of mass of the plaque moves from S0 to S2. The process can be considered that the center of mass S0 of the plaque first moves to S1, that is, the overall position of the plaque moves from S32 to S37, and then the center of mass moves from S1 to S2, that is, the overall position of the plaque moves from S37 to S36. Therefore, the displacement vector S34 of the actual plaque deformation is expressed as follows:

[0073] The vector addition diagram is shown in Figure 5

[0074] The coarse strain image frame in the coarse and fine strain image frames can correspond to the longitudinal strain image frame, that is, the displacement from S0 to S1 The fine strain image frame can correspond to the lateral strain image frame, that is, the displacement from S1 to S2

[0075] As shown in FIG4 , S40 is the decomposed longitudinal strain image frame, defined as PH(t, i, j), S41 is the decomposed transverse strain image frame, defined as PL(t, i, j), S42 is the longitudinal direction, and S43 is the transverse direction;

[0076] As shown in step S13 of FIG1 , the coarse and fine strain image frames of the tissue are calculated and acquired. The coarse strain image is calculated using the phase iterative difference method. A calculation formula is as follows:

[0077] in:

[0078] w0 is the angular frequency of the deformation; arg: represents the phase angle; H0∈[-H:1:H]; p(t,i,j) represents the displacement of the pixel in the i-th row and j-th column in the frame at time t;

[0079] As shown in step S13 of FIG1 , the coarse and fine strain image frames of the tissue are calculated and acquired, wherein the fine strain image is calculated using the following formula:

[0080] pl(t,i,j)=max j {R(t,i,j)}

[0081] Where: L0∈[-L:1:L], I: is a fixed row on the image, max j {} is the maximum value function in the j direction, l represents the variable in the interval [-L0, L0]; dl represents the integral variable;

[0082] As shown in step S14 of Figure 1, the coarse strain image frame PH(t,i,j) and the fine strain image frame PL(t,i,j) are fused and matched to obtain the precise strain image frame. A fusion method can directly use the displacement information of pl(t,i,j) to match the p in step 5. t-1 (t,i,j) displacement frame vector calibration, as in step 3 The formula shown is calculated Later In the vector direction, the formula for calculating PH(t,i,j) in step 5 is used to obtain the precise strain force image frame P(t,i,j);

[0083] For the precise strain image frame in step 7, one or more sampling points can be selected in the region of interest, i.e., i and j can be arbitrarily selected to obtain the precise strain value P(t) that changes with time, as shown in S64 in FIG6 , where the abscissa is the time axis S63 and the ordinate is parameter 0, i.e., S60. Parameter 0 can be defined as the precise strain value P(t). The displayed curve is S64, and S62 is the curve 0 in the icon drawn with a solid line or curve 1 drawn with a dotted line.

[0084] As shown in step S15 of FIG1 , the precise strain image P(t,i,j) is counted for the number of times a certain strain pair PV(t,c(v),i,j) has a certain strain value co(v,n). The statistical results are processed to form a feature image FI(t,n,n). The specific formula is as follows:

[0085] FI(t,n,n)=SF{PV(t,c(v),i,j)==co(v,n)} / Factor

[0086] Where: v represents the pixel sequence in the pixel pair, v = 1, 2, 3...; N is the maximum strain force, co(v,n) is the prior input value; SF{} is the statistical frequency function of strain force co(v,n); n∈[0,N-1]; Factor is the normalization factor.

[0087] As shown in FIG1 , S16 performs feature analysis on the feature image FI(t,n,n) to obtain various types of mathematical or image feature indicators, such as parameter 1, parameter 2, parameter 3, parameter 4, etc., as shown in S61 in FIG6 . Without loss of generality, this patent can calculate various parameters such as the energy Power of the plaque tissue based on FI(t,n,n). One energy Power formula is as follows:

[0088] Power=sum{FI(t,n,n)*FI(t,n,n)},n∈[0,N-1]

[0089] Where sum{} is the sum function;

[0090] As shown in step S17 of FIG1 , the final strain image and characteristic index are comprehensively displayed and output to the display for user diagnosis. As shown in FIG7 , S70 and S71 are dual real-time displayed images, S70 is the grayscale image area, S71 is a composite image of the grayscale image S73 and the precise strain image S72, where the image of S73 is consistent with the grayscale image of S70, and S74 is a characteristic highlighting area for displaying various parameters such as the index and charts, as shown in FIG6 .

[0091] This patent enables quantitative measurement and qualitative analysis of plaque within arterial vessels. This patent utilizes elastic imaging technology from medical ultrasound equipment to improve the accuracy of plaque detection within arterial vessels while minimizing data volume. The patent also calculates various plaque characteristic indicators based on plaque strain information, conveniently and quickly transmitting this information to the backend for processing and display on a monitor for user diagnosis.

[0092] The arterial plaque characteristic detection device based on elastic imaging provided by the present invention is applied to medical color Doppler ultrasound equipment, comprising:

[0093] Select the tissue region of interest module to automatically obtain the manually obtained carotid artery plaque area;

[0094] Transmitter and receiver modules are used to obtain ultrasonic signals of tissue deformation;

[0095] A coarse and fine strain image frame calculation module is used to calculate multiple frames of coarse and fine strain images;

[0096] A precise strain image frame calculation module, used to calculate precise strain image frames;

[0097] A feature image calculation module is used to calculate the number of times a certain pixel has a certain strain value in response to the strain;

[0098] Parameter acquisition module, used to extract mathematical or image feature indicators from feature images;

[0099] The display module is used to display the final strain image and characteristic index values.

[0100] As shown in FIG2 , in this embodiment, a device for detecting arterial plaque characteristics based on elastic imaging of a medical color Doppler ultrasound device includes:

[0101] A tissue region of interest selection module S20 is used to select an effective plaque diagnosis area;

[0102] Transmitting and receiving module S21, used to obtain ultrasonic echo data including tissue deformation;

[0103] A coarse and fine strain image frame calculation module S22 is used to calculate longitudinal and transverse strain image frame data of the user's area of ​​interest;

[0104] The precise strain image frame calculation module S23 is used to fuse the coarse and fine strain images to obtain a precise strain image;

[0105] A characteristic image calculation module S24 is used to obtain a data frame for calculating plaque characteristic parameters;

[0106] Parameter acquisition module S25, used to calculate the index value or chart of plaque characteristics;

[0107] Display module S26, used to display strain images and various parameter indicators of plaque characteristics;

[0108] The present invention provides a computer-storable medium, which is used to execute the above-mentioned arterial vascular plaque characteristic detection method based on elastic imaging.

[0109] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. An arterial vascular plaque characteristic detection method based on elastic imaging, applied to medical color Doppler ultrasound equipment, characterized in that: include: Acquiring a tissue region of interest containing plaque features within an arterial vessel; Transmit and receive ultrasonic signals in the tissue area of ​​interest according to the tissue movement frequency; Obtaining coarse and fine strain force image frames by calculating ultrasonic signals; The coarse and fine strain image frames are fused and matched to obtain the precise strain image frame, and the precise strain image frame is obtained; Counting the number of times each pixel in the precise strain image frame has a strain value, and performing image processing on the statistical results to obtain a characteristic image; Perform feature analysis on the feature image, obtain feature analysis results, and output them for display.

2. The arterial vascular plaque characteristic detection method based on elastic imaging as claimed in claim 1, characterized in that: When acquiring a tissue region of interest containing plaque features within an arterial vessel, the ultrasound software is controlled to automatically locate the arterial plaque region of the current image as the tissue region of interest.

3. The arterial vascular plaque characteristic detection method based on elastic imaging as claimed in claim 2, characterized in that: The user may manually select a region of interest as the tissue region of interest.

4. The method for detecting arterial vascular plaque characteristics based on elastic imaging as claimed in claim 1, characterized in that: The scanning sampling frequency of transmitting and receiving ultrasonic signals is more than twice the frequency of tissue movement.

5. The arterial vascular plaque characteristic detection method based on elastic imaging as claimed in claim 4, characterized in that: The scanning sampling frequency of transmitting and receiving ultrasonic signals can be adjusted by the user.

6. The arterial vascular plaque characteristic detection method based on elastic imaging as claimed in claim 1, characterized in that: The ultrasonic signal can be received by using two or more different methods to calculate and obtain coarse and fine strain image frames, wherein the strain image method calculated by a high-precision slow method is a fine image, and the strain image method calculated by a low-precision fast method is a coarse image.

7. The arterial vascular plaque characteristic detection method based on elastic imaging as claimed in claim 1, characterized in that: The coarse and fine strain image frames are fused and matched, and the fusion matching includes the matching of the positions of each pixel point between the coarse and fine image frames and the superposition of the strain force.

8. The method for detecting arterial vascular plaque characteristics based on elastic imaging as claimed in claim 1, characterized in that: The time-varying pattern of any pixel in the precise strain image frame can be obtained. The precise strain image frame is a continuous frame in time. The system can automatically or manually select quantitative measurement points on the continuous frame. The ultrasound software can display and draw the numerical value, curve or chart of the quantitative measurement point changing with time.

9. The method for detecting arterial vascular plaque characteristics based on elastic imaging as claimed in claim 1, characterized in that: The number of times a certain pixel has a certain strain value for the precise strain image containing the plaque is counted, and the statistical process can be calculated in a frozen state of the image or in a real-time state.

10. The arterial vascular plaque characteristic detection method based on elastic imaging as claimed in claim 1, characterized in that: The strain force of a certain pixel includes the strain force of a pixel point set composed of one or two or more pixels.

11. The method for detecting arterial vascular plaque characteristics based on elastic imaging according to claim 1, characterized in that: The statistical results are processed to form a characteristic image, wherein the data processing is a process of normalizing the statistical times or recalculating the probability distribution.

12. The arterial vascular plaque characteristic detection method based on elastic imaging as claimed in claim 1, characterized in that: Feature analysis of feature images can obtain various types of mathematical or image feature indicators. Feature analysis is based on normalized or re-probability distributed feature images. Mathematical or image feature indicators include but are not limited to brightness, energy, contrast, uniformity, non-uniformity, homogeneity, non-homogeneity, correlation, non-correlation and other Chinese or English abbreviations.

13. The arterial vascular plaque characteristic detection method based on elastic imaging as claimed in claim 1, characterized in that: The final strain image and characteristic index are comprehensively displayed and output to a display for user diagnosis, wherein the characteristic index can be displayed dynamically or statically at any position of the display in the form of numerical values, charts, curves but not limited to such forms.

14. An arterial vascular plaque characteristic detection device based on elastic imaging, applied to medical color Doppler ultrasound equipment, characterized in that: include: Select the tissue region of interest module to automatically obtain the manually obtained carotid artery plaque area; A transmitting and receiving module is used to obtain ultrasonic signals of tissue deformation; A coarse and fine strain force image frame calculation module, used to calculate coarse and fine strain force multi-frame images; A precise strain image frame calculation module, used to calculate precise strain image frames; A characteristic image calculation module is used to calculate the number of times a certain pixel has a certain strain value for the strain; A parameter acquisition module is used to extract mathematical or image feature indicators from feature images; The display module is used to display the final strain image and characteristic index value.

15. A computer storable medium, characterized in that: The computer storable medium is used to execute the arterial vascular plaque characteristic detection method based on elastic imaging as claimed in any one of claims 1 to 13.

Images