Data processing device, data processing method, and data processing program product

Abstract

Provided are a data processing device, a data processing method, and a program with which it is possible to easily evaluate an embryo obtained by in vitro fertilization. A data processing device is provided with: an acquisition unit that acquires microscope image data of an embryo obtained by in vitro fertilization, and request statement data indicating an evaluation request for the embryo reflected in the microscope image data; an evaluation unit that inputs the microscope image data and the request statement data to a data generation model; obtaining evaluation data of an embryo reflected in the microscope image data output from the data generation model, wherein the evaluation data corresponds to the request statement data; and an output unit that outputs evaluation data of the embryo.

Description

Technical Field

[0001] The technology disclosed herein relates to a data processing apparatus, a data processing method, and a data processing program product. Background Technology

[0002] Patent document 1 discloses an apparatus, method, and system for image-based classification of human embryonic cells.

[0003] Patent document 2 discloses an artificial intelligence (AI) computing system for generating embryo viability scores based on a single image of an embryo, in order to help select embryos for implantation in in vitro fertilization (IVF).

[0004] Patent Document 3 discloses a system for predicting the viability of one or more embryos. The system disclosed in Patent Document 3 may include: receiving a single image of an embryo via a real-time communication link with an image capture device; and classifying the single image using at least one convolutional neural network to generate a viability score associated with the embryo.

[0005] Existing technical documents

[0006] Patent documents

[0007] Patent Document 1: Japanese Patent Application Publication No. 2022-087297

[0008] Patent Document 2: Japanese Patent Publication No. 2022-528961

[0009] Patent Document 3: Japanese Patent Publication No. 2024-513659 Summary of the Invention

[0010] The problem that the invention aims to solve

[0011] However, when evaluating embryos obtained through in-vitro fertilization in a system, it is preferable to perform the evaluation as desired by the user. For example, if the user desires a predicted implantation rate, the system preferably outputs the implantation rate in real time. Similarly, if the user wants to know the embryo's grade, the system preferably outputs the embryo's grade in real time. Furthermore, if supplementary information related to the embryo being evaluated exists, the system preferably evaluates the embryo while also taking this supplementary information into account. Such a system requires interaction with the user.

[0012] Regarding this point, there is still room for improvement in the existing technology. Specifically, the input data in the technologies disclosed in Patent Documents 1-3 is only image data of the embryo, and it cannot process sentence data input from the user. Therefore, the problem with the technologies disclosed in Patent Documents 1-3 is that it is impossible to evaluate the embryo obtained through in vitro fertilization while interacting with the user.

[0013] Methods for solving problems

[0014] The first aspect of the technology disclosed herein relates to a data processing apparatus comprising: an acquisition unit that acquires microscope image data of an embryo obtained through in vitro fertilization and request statement data, the request statement data representing an evaluation request for the embryo reflected in the microscope image data; an evaluation unit that inputs the microscope image data and the request statement data to a data generation model and acquires evaluation data of the embryo reflected in the microscope image data, output from the data generation model, the evaluation data corresponding to the request statement data; and an output unit that outputs the evaluation data of the embryo.

[0015] The second aspect of the technology disclosed herein is a data processing method comprising: acquiring microscopic image data of an embryo obtained through in vitro fertilization and request statement data, wherein the request statement data represents an evaluation request of the embryo reflected in the microscopic image data; inputting the microscopic image data and the request statement data into a data generation model; acquiring evaluation data of the embryo reflected in the microscopic image data output from the data generation model, wherein the evaluation data corresponds to the request statement data; and outputting the evaluation data of the embryo.

[0016] The third aspect of the technology disclosed herein is a program product that stores a program for causing a computer to perform processing, the processing including: acquiring microscopic image data of an embryo obtained through in vitro fertilization and request statement data, the request statement data representing an evaluation request of the embryo reflected in the microscopic image data; inputting the microscopic image data and the request statement data into a data generation model; acquiring evaluation data of the embryo reflected in the microscopic image data output from the data generation model, the evaluation data corresponding to the request statement data; and outputting the evaluation data of the embryo. Attached Figure Description

[0017] Figure 1 This is a schematic diagram illustrating an example of the structure of a data processing system.

[0018] Figure 2 This is a schematic diagram illustrating an example of the main functions of a data processing device and a user terminal.

[0019] Figure 3 This is a diagram used to illustrate an embryo obtained through in-vitro fertilization.

[0020] Figure 4 The functional structure of the determination processing unit of the data processing apparatus is shown in summary.

[0021] Figure 5A This is an example of a screen displayed on a user's terminal monitor.

[0022] Figure 5B This is an example of a screen displayed on a user's terminal monitor.

[0023] Figure 5C This is an example of a screen displayed on a user's terminal monitor.

[0024] Figure 6 This section provides a summary of an example of the action flow of a deterministic process performed by a data processing device. Detailed Implementation

[0025] Hereinafter, with reference to the accompanying drawings, an example of an implementation of the data processing apparatus, data processing method, and program involved in the present disclosure will be described.

[0026] First, the terminology used in the following explanation will be explained.

[0027] In the following embodiments, the processor (hereinafter referred to as "processor") marked with reference numerals can be a single computing device or a combination of multiple computing devices. Furthermore, a processor can be a single computing device or a combination of multiple computing devices. Examples of computing devices include a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), a General-Purpose Computing on Graphics Processing Units (GPGPU), or an Accelerated Processing Unit (APU).

[0028] In the following implementation, the Random Access Memory (RAM) with reference numerals is a memory for temporary storage of information, which is used by the processor as working memory.

[0029] In the following embodiments, the storage device (with reference numerals) is one or more non-volatile storage devices that store various programs and parameters. Examples of non-volatile storage devices include flash memory (SSD), disk drives (e.g., hard disks), or magnetic tape.

[0030] In the following embodiments, the communication interface (I / F) with reference numerals is an interface that includes a communication processor and an antenna, etc. The communication I / F is responsible for communication between multiple computers. As examples of communication standards applicable to the communication I / F, wireless communication standards include 5th Generation Mobile Communication System (5G), Wi-Fi (registered trademark), or Bluetooth (registered trademark).

[0031] In the following embodiments, "A and / or B" has the same meaning as "at least one of A and B". That is, "A and / or B" can mean only A, only B, or a combination of A and B. In addition, in this specification, when three or more items are described by connecting them with "and / or", the same concept as "A and / or B" also applies.

[0032] exist Figure 1 The diagram shows an example of the structure of the data processing system 10 according to the embodiment.

[0033] like Figure 1 As shown, the data processing system 10 includes a data processing device 12 and a user terminal 14. An example of the data processing device 12 is a server. An example of the user terminal 14 includes a personal computer or a smartphone. In this embodiment, the data processing device 12 is an example of a "data processing device" according to the technology of this disclosure, and the user terminal 14 is an example of a "terminal" according to the technology of this disclosure.

[0034] The data processing apparatus 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" as understood in this disclosure. The computer 22 includes a processor 28, RAM 30, and storage device 32. The processor 28, RAM 30, and storage device 32 are connected to a bus 34. Additionally, the database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a wide area network (WAN) and / or a local area network (LAN).

[0035] User terminal 14 includes a computer 36, a receiving device 38, an output device 40, a camera 42, and a communication I / F 44. Computer 36 includes a processor 46, RAM 48, and storage device 50. The processor 46, RAM 48, and storage device 50 are connected to a bus 52. Additionally, the receiving device 38, output device 40, and camera 42 are also connected to the bus 52.

[0036] The receiving device 38 includes a touch panel 38A and a microphone 38B, etc., to receive user input. The touch panel 38A receives user input generated by the contact of an indicator (e.g., a pen or finger). The microphone 38B receives user input generated by sound by detecting the user's voice. The control unit 46A sends the data representing the user input received through the touch panel 38A and the microphone 38B to the data processing unit 12. In the data processing unit 12, the determination processing unit 290 acquires the data representing the user input.

[0037] The output device 40 includes a display 40A and a speaker 40B, which output data to the person 20 in a perceptible form (e.g., sound and / or text). The display 40A displays visual information such as text and images according to instructions from the processor 46. The speaker 40B outputs sound according to instructions from the processor 46. The camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element including a complementary metal-oxide-semiconductor (CMOS) image sensor or a charge-coupled device (CCD) image sensor.

[0038] Communication I / F 44 is connected to network 54. Communication I / F 44 and communication I / F 26 are responsible for sending and receiving various information between processor 46 and processor 28 via network 54.

[0039] exist Figure 2 The image shows an example of the main functions of the data processing device 12 and the user terminal 14.

[0040] like Figure 2As shown, in the data processing apparatus 12, the determination process is performed by the processor 28. The storage device 32 stores a determination processing program 56. The determination processing program 56 is an example of a "program" according to the technology disclosed herein. The processor 28 reads the determination processing program 56 from the storage device 32 and executes the read determination processing program 56 on the RAM 30. The determination process is implemented by operating as a determination processing unit 290 based on the determination processing program 56 executed by the processor 28 on the RAM 30.

[0041] The storage device 32 stores a data generation model 58. The data generation model 58 is used by the determination processing unit 290.

[0042] In user terminal 14, receive / output processing is performed by processor 46. Receive / output program 62 is stored in storage device 50. Receive / output program 62 is used in combination with determination processing program 56 via data processing system 10. Processor 46 reads receive / output program 62 from storage device 50 and executes the read receive / output program 62 in RAM 48. Receive / output processing is implemented by operating control unit 46A according to the receive / output program 62 executed by processor 46 in RAM 48.

[0043] Next, the data processing device 12 will describe the processing of the determination processing unit 290 during the determination processing of the embryo obtained by in vitro fertilization.

[0044] Figure 3 This is a diagram used to illustrate embryos obtained through in-vitro fertilization. For example... Figure 3 As shown, embryos obtained through in-vitro fertilization develop into morulae through 2-cell, 4-cell, and 8-cell stages, and then into blastocysts. The blastocysts are then transferred back to the uterus to promote implantation. (Regarding...) Figure 3 The embryos shown are visually examined by doctors, who select only those in good condition for transfer back to the uterus. However, the evaluation criteria for embryos obtained through visual examination are vague, and evaluation results may vary. For example, one doctor may evaluate an embryo shown in a microscope image as "good," while another doctor may evaluate the same embryo as "poor."

[0045] Therefore, in this embodiment, the data generation model 58, described later, is used to evaluate embryos obtained through in vitro fertilization. This reduces the discrepancies in evaluation results obtained through visual inspection. Furthermore, as described later, the data generation model 58 can also process statement data, thus enabling the evaluation of the embryo as desired by the user while interacting with the user.

[0046] like Figure 4 As shown, the determination processing unit 290 includes an acquisition unit 292, an evaluation unit 294, and an output unit 296.

[0047] The acquisition unit 292 acquires user input received on the user terminal 14. Specifically, it acquires at least one type of data from the user, including text, voice, and images, received on the user terminal 14. It should be noted that the user in this embodiment is, for example, a doctor.

[0048] Specifically, the acquisition unit 292 acquires microscope image data of an embryo obtained through in vitro fertilization, input by the user, and request statement data, which represents an evaluation request for the embryo reflected in the microscope image data.

[0049] Figures 5A to 5C This is an example of a screen displayed on the display 40A of the user terminal 14. The acquisition unit 292 of the data processing device 12 displays on the display 40A of the user terminal 14. Figures 5A to 5C The image shown.

[0050] exist Figure 5A The left side of the screen shows the source data used for learning data generation model 58. Figure 5A On the right side of the screen shown, there is a field for inputting the microscope image data of the embryo being analyzed. Figure 5A (The "Upload Image" section). It should be noted that in... Figure 5A On the right side of the screen shown, there is a field for inputting supplementary information related to the embryo being analyzed. It should be noted that this field can accept user-inputted statements.

[0051] like Figure 5B As shown, when the microscope image data of the object to be analyzed is input, the request statement data such as "Analyze the implantation rate of this embryo" is entered in the input supplementary information field, and the "Start Analysis" button is clicked, the acquisition unit 292 acquires the microscope image data and the request statement data. The statement data such as "Analyze the implantation rate of this embryo" also serves as a prompt to the data generation model 58.

[0052] The evaluation unit 294 performs a determination process using the data generation model 58. Specifically, the evaluation unit 294 inputs the microscope image data and request statement data acquired by the acquisition unit 292 into the data generation model 58 to obtain evaluation data of the embryo as the generation result. More specifically, the evaluation unit 294 acquires the evaluation data of the embryo reflected in the microscope image data output from the data generation model 58, and this evaluation data corresponds to the request statement data.

[0053] The output unit 296 sends evaluation data as the result of the determination process to the user terminal 14. In the user terminal 14, the control unit 46A outputs the evaluation data as the result of the determination process to the display 40A of the output device 40.

[0054] Figure 5C This is a diagram showing an example of evaluation data displayed on monitor 40A. Figure 5C The evaluation data shown includes statement data representing the implantation rate, embryo grade, and embryo evaluation as reflected in the microscopic image data.

[0055] Data generation model 58 is a so-called generative system AI (Artificial Intelligence). As an example of data generation model 58, ChatGPT (Internet Search) is cited.<URL:https: / / openai.com / blog / chatgpt> Gemini (Internet Search)<URL:https: / / gemini.google.com / ?hl=ja> This refers to an AI-generated system, such as a data generation model 58. The data generation model 58 is obtained through deep learning of a neural network. Input to the data generation model 58 includes prompts containing instructions, as well as inference data such as sound data representing sound, text data representing text, and image data representing images. The data generation model 58 infers the input inference data based on the instructions shown in the prompts, and outputs the inference results in the form of sound data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0056] It should be noted that users can also click... Figure 5A The "Additional Learning" button on the screen indicates that the data generation model 58 is performing additional learning and fine-tuning. In this case, for example, the acquisition unit 292 acquires learning data as a combination of learning microscope image data and learning evaluation data corresponding to the learning microscope image data. The learning microscope image data reflects an embryo different from the embryo reflected in the microscope image data of the analysis object, and reflects an embryo provided by a patient who provided the embryo reflected in the microscope image data of the analysis object.

[0057] Next, the evaluation unit 294, based on the learning data, uses a known machine learning algorithm to fine-tune the data generation model 58, thereby generating a new data generation model 58 for the patient. Then, the evaluation unit 294 inputs the microscope image data of the analysis object into the new data generation model 58 for the patient, and obtains evaluation data of the embryo reflected in the microscope image data, output from the new data generation model 58 for the patient. In this way, the evaluation unit 294 can generate evaluation data suitable for the patient by fine-tuning the data generation model 58 for the patient.

[0058] In addition, users can Figure 5A In the "Supplementary Information" field shown on the screen, you can enter various statements as prompts or requests. For example, you can also enter statements related to the patient's specific circumstances, such as the embryos being analyzed.

[0059] In addition, users can also Figure 5A In the "Upload Image" field of the displayed screen, instead of uploading individual microscope image data, time-series data (growth data) representing the development process of an embryo is uploaded. In this case, the acquisition unit 292 acquires the growth data, which is time-series data of microscope image data reflecting the development process of an embryo provided by the same patient, and specifically time-series data of microscope image data representing the development process of an embryo. The evaluation unit 294 inputs the growth data acquired by the acquisition unit 292 and the request statement data into the data generation model 58. The data generation model 58 outputs evaluation data corresponding to the input growth data. The evaluation unit 294 acquires the evaluation data output from the data generation model 58. It should be noted that the evaluation unit 294 may also use known image processing techniques to extract feature data from the microscope image data and input this feature data into the data generation model 58.

[0060] In addition, Figure 5C The example shown illustrates evaluation data that includes statements indicating the implantation rate, embryo grade, and embryo evaluation of embryos reflected in microscope image data. However, it is also possible to output only the evaluation data corresponding to the request statement data. For example, if the request statement data such as "Please calculate only the implantation rate" is input, the data generation model 58 outputs only the implantation rate as evaluation data. Similarly, if the request statement data such as "Please output only the grade" is input, the data generation model 58 outputs only the embryo grade as evaluation data. Thus, when evaluating embryos reflected in microscope image data, the evaluation desired by the user can be performed more flexibly by using the data generation model 58, which is capable of processing statement data.

[0061] Next, the function of the data processing system 10 will be explained.

[0062] Reference Figure 6 And it illustrates an example of a defined processing flow. It should be noted that... Figure 6 The process shown is an example of the "data processing method" involved in the technology of this disclosure.

[0063] In step S300, the acquisition unit 292 acquires the microscope image data of the embryo of the analysis object and the request statement data input by the user.

[0064] In step S301, the evaluation unit 294 inputs the microscopic image data of the embryo and the request statement data obtained in step S300 into the data generation model 58.

[0065] In step S302, the evaluation unit 294 acquires evaluation data of the embryo reflected in the microscope image data output from the data generation model 58, and the evaluation data corresponds to the request statement data.

[0066] In step S303, the output unit 296 outputs evaluation data to the user terminal 14 and ends the determination process.

[0067] Thus, the data processing apparatus of this embodiment acquires microscopic image data of an embryo obtained through in vitro fertilization and request statement data, which represents an evaluation request for the embryo reflected in the microscopic image data. The data processing apparatus inputs the microscopic image data and the request statement data into a data generation model, and acquires evaluation data of the embryo reflected in the microscopic image data, output from the data generation model, which corresponds to the request statement data. The data processing apparatus outputs the evaluation data of the embryo. Therefore, the evaluation of embryos obtained through in vitro fertilization can be performed conveniently. Furthermore, the discrepancies in evaluation results of embryos obtained through visual examination by doctors or others can be reduced. Moreover, by using a data generation model capable of processing statement data, the embryo can be evaluated in the manner desired by the user while interacting with the user.

[0068] The data processing apparatus of this embodiment utilizes AI (Artificial Intelligence) technology for evaluating embryos conceived through in-vitro fertilization. In this embodiment, the AI ​​analyzes microscopic images or growth data of embryos to identify embryos with a high probability of implantation. This leads to improved treatment success rates and an increased birth rate of healthy babies. Furthermore, it reduces the burden on infertile patients, improves their satisfaction, and promotes efficient use of medical resources. Additionally, it enables the provision of personalized treatment plans and support to patients.

[0069] The above description mainly focuses on the functions of the data processing device 290, and describes the system involved in this disclosure. However, the system involved in this disclosure is not limited to being installed on a server. The system involved in this disclosure can also be installed as a general information processing system. For example, this disclosure can also be installed as a software program that operates on a personal computer, or an application program that operates on a smartphone, etc. The method involved in this disclosure can also be provided to users in the form of Software as a Service (SaaS).

[0070] In the above embodiments, an example of determining the process is given by one computer 22, but the technology disclosed herein is not limited to this, and may also include multiple computers of computer 22 performing distributed processing of the determination process.

[0071] In the above embodiments, an example of storing the determination process 56 in the storage device 32 has been given and described, but the technology disclosed herein is not limited thereto. For example, the determination process 56 may also be stored in a portable computer-readable non-transitory storage medium such as a Universal Serial Bus (USB) memory. The determination process 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 performs determination processing according to the determination process 56.

[0072] Alternatively, a determination process 56 may be stored in a storage device such as a server connected to the data processing device 12 via the network 54, and the determination process 56 corresponding to the request of the data processing device 12 may be downloaded and installed into the computer 22.

[0073] It should be noted that the entire contents of the determination process 56 may be stored in a storage device such as a server connected to the data processing device 12 via the network 54, or it may be unnecessary to store the entire contents of the determination process 56 in the storage device 32, but only a part of the determination process 56 may be stored.

[0074] As hardware resources for performing deterministic processing, various processors can be used, as shown below. For example, a CPU is a general-purpose processor that functions as a hardware resource to perform deterministic processing by executing software, i.e., a program. Other examples of processors include field-programmable gate arrays (FPGAs), programmable logic devices (PLDs), or processors with special-purpose circuitry (ASICs) designed specifically for performing particular processing. Regardless of the processor, it has built-in or connected memory, and it performs deterministic processing using this memory.

[0075] The hardware resources for performing deterministic processing can consist of one of these various processors, or a combination of two or more processors of the same or different types (e.g., a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Alternatively, the hardware resources for performing deterministic processing can also be a single processor.

[0076] As an example of a single processor, the first approach is to combine one or more CPUs with software to form a single processor, which functions as a hardware resource for performing deterministic processing. The second approach, exemplified by System-on-a-chip (SoC), uses a processor that implements the overall functionality of a system containing multiple hardware resources for performing deterministic processing on a single IC chip. In this way, deterministic processing is implemented using one or more of the aforementioned processors as hardware resources.

[0077] Furthermore, as for the hardware architecture of these various processors, more specifically, circuits composed of semiconductor elements and other circuit components can be used. Also, the aforementioned process is merely one example. Therefore, it goes without saying that unnecessary steps can be deleted, new steps can be added, or the processing order can be adjusted without departing from the main point.

[0078] The foregoing descriptions and illustrations are detailed explanations of the parts covered by this disclosure, and are merely one example of the technology disclosed herein. For example, the descriptions related to the structure, function, role, and effect described above are examples of the structure, function, role, and effect of the parts covered by this disclosure. Therefore, without departing from the spirit of this disclosure, unnecessary parts may be deleted, new elements added, or replacements may be made with the foregoing descriptions and illustrations; this is self-evident. Furthermore, to avoid complexity and facilitate understanding of the parts covered by this disclosure, descriptions related to common technical knowledge that enables the implementation of this disclosure and does not require special explanation have been omitted from the foregoing descriptions and illustrations.

[0079] All documents, patent applications and technical standards described in this specification are incorporated herein by reference to the same extent as each document, patent application and technical standard is incorporated by reference in its specific and individual case.

[0080] [Explanation of the labels in the attached diagram]

[0081] 10 Data Processing System

[0082] 12 Data processing device

[0083] 14 User Terminals

[0084] 290 Determining the Processing Department

[0085] 292 Acquisition Department

[0086] 294 Evaluation Department

[0087] 296 Output Section

Claims

1. A data processing apparatus, wherein, The acquisition unit acquires microscopic image data of an embryo obtained through in vitro fertilization and request statement data, wherein the request statement data represents an evaluation request for the embryo reflected in the microscopic image data. An evaluation unit inputs the microscope image data and the request statement data into a data generation model, and obtains evaluation data of the embryo reflected in the microscope image data output by the data generation model, wherein the evaluation data corresponds to the request statement data. as well as The output unit outputs the evaluation data of the embryo.

2. The data processing apparatus according to claim 1, wherein, The acquisition unit acquires learning data, which is a combination of learning microscope image data and learning evaluation data corresponding to the learning microscope image data. The learning microscope image data reflects an embryo different from the embryo reflected in the microscope image data, and reflects an embryo provided by a patient who provided the embryo reflected in the microscope image data. The evaluation department fine-tunes the data generation model based on the learning data, thereby generating a new data generation model for the patient. The microscope image data is input into a new data generation model for the patient to obtain evaluation data of the embryo reflected in the microscope image data, output from the new data generation model for the patient.

3. The data processing apparatus according to claim 1 or 2, wherein, The evaluation data includes at least one of the following statement data: the implantation rate of the embryo reflected in the microscope image data, the grade of the embryo, and the evaluation of the embryo.

4. The data processing apparatus according to claim 1 or 2, wherein, The acquisition unit acquires growth data, which is time-series data of microscopic image data reflecting the growth process of an embryo provided by the same patient. The evaluation unit inputs the growth data and the request statement data into the data generation model and obtains the evaluation data output from the data generation model.

5. A data processing method, wherein, The data processing method is performed by a computer, and the processing includes: Acquire microscopic image data of embryos obtained through in vitro fertilization, and request statement data, wherein the request statement data represents an evaluation request for the embryo reflected in the microscopic image data. The microscope image data and the request statement data are input into the data generation model to obtain evaluation data of the embryo reflected in the microscope image data, which is output from the data generation model. The evaluation data corresponds to the request statement data. Output the evaluation data of the embryo.

6. A program product, wherein, The program product stores a program for causing a computer to perform processing, the processing including: Acquire microscopic image data of embryos obtained through in vitro fertilization, and request statement data, wherein the request statement data represents an evaluation request for the embryo reflected in the microscopic image data. The microscope image data and the request statement data are input into the data generation model to obtain evaluation data of the embryo reflected in the microscope image data, which is output from the data generation model. The evaluation data corresponds to the request statement data. Output the evaluation data of the embryo.

Images