JP2016171900A - Endoscope apparatus and program execution method thereof - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an endoscope apparatus including a plurality of arithmetic units capable of executing parallel processing in which processing of a program is appropriately allocated to the plurality of arithmetic units that execute the processing.SOLUTION: An endoscope apparatus includes a plurality of arithmetic units, a first acquisition part for acquiring usage rates of the plurality of arithmetic units, a storage part for storing the usage rates of a plurality of arithmetic units used for the execution of a program for each program executed in the past, and a determination part for determining a program to be executed next from programs awaiting execution based on the usage rate acquired by the first acquisition part and the usage rate for each program stored in the storage part.SELECTED DRAWING: Figure 2

Description

  The present invention relates to an endoscope apparatus that processes a video signal acquired by an endoscope and a program execution method thereof.

  Conventionally, a GPU (Graphics Processing Unit) is known in the field of processing video signals. The GPU has a configuration including a plurality of CUs (Compute Units) including, for example, a plurality of VALUs (Vector Arithmetic Logic Units), and can perform video processing and the like at a very high speed. In recent years, GPUs with a mechanism for parallel processing of multiple kernels (programs running on GPUs) called CKE (Concurrent Kernel Execution) are common, and advanced parallel processing that fully utilizes all CUs It has become feasible.

  On the other hand, as to how to divide kernel processing on the GPU and how to divide and execute the divided processing on the CU, there are many parts left to the kernel compiler or scheduler on the GPU. It depends on the state of the GPU (hardware resource usage, etc.) when executing the kernel. This is because even if the kernel executed by the GPU is the same, the time required for the GPU to complete the execution of the kernel after it is instructed to the GPU is different each time. Suggests that time cannot be predicted.

  For this reason, it is not possible to easily adopt a GPU in an endoscope apparatus in the medical field that has a limitation in processing time. If a GPU is used, a mechanism for appropriately dividing kernel processing and appropriately distributing the divided processing to CUs is required.

The following devices are also known in the field of processing video signals.
For example, in an image forming apparatus having a plurality of drawing units that respectively draw a plurality of drawing areas of image data based on a drawing command, the load of drawing processing on each of the plurality of drawing areas when the current image data is drawn The amount is calculated, and based on the drawing command, a change in the drawing processing load amount for each of the plurality of drawing regions when the next image data is drawn is predicted. Based on the calculation result and the prediction result, It is known that when the next image data is drawn, the amount of drawing processing applied to each of a plurality of drawing areas is calculated, and processing is assigned to a plurality of drawing means based on the calculation result. (See, for example, Patent Document 1).

JP 2013-127664 A

  In light of the above situation, the present invention provides an endoscope apparatus that includes a plurality of arithmetic units capable of parallel processing, and can appropriately distribute and execute program processing among the plurality of arithmetic units. An object of the present invention is to provide an endoscope apparatus and a program execution method thereof.

  The first aspect of the present invention was used to execute a plurality of arithmetic units, a first acquisition unit that acquires the usage rate of the plurality of arithmetic units, and each program executed in the past. Based on the storage unit that stores the usage rates of the plurality of arithmetic units, the usage rate acquired by the first acquisition unit, and the usage rate for each program stored in the storage unit, An endoscope apparatus comprising: a determining unit that determines a program to be executed next from among programs.

  According to a second aspect of the present invention, in the first aspect, when the program is executed, the second aspect further includes a second acquisition unit that acquires usage rates of the plurality of arithmetic units used to execute the program. An endoscope apparatus in which the usage rate acquired by the second acquisition unit is stored in the storage unit is provided.

  According to a third aspect of the present invention, in the first or second aspect, the determining unit is a program that can immediately start execution, and that has the earliest waiting order for execution, An endoscope apparatus that is determined as a program to be executed next is provided.

  According to a fourth aspect of the present invention, in any one of the first to third aspects, the determination unit includes a program that must be executed within a predetermined time in the execution waiting program. There is provided an endoscope apparatus in which the program is prioritized and determined as a program to be executed next.

  A fifth aspect of the present invention provides an endoscope apparatus according to any one of the first to fourth aspects, wherein the plurality of arithmetic units are a plurality of arithmetic units included in a GPU (Graphics Processing Unit). To do.

  According to a sixth aspect of the present invention, there is provided a program execution method for an endoscope apparatus including a plurality of arithmetic units, wherein the usage rates of the plurality of arithmetic units are acquired, and the usage rates of the acquired arithmetic units are acquired. Determining a program to be executed next from among the programs waiting to be executed based on each of the programs executed in the past and the usage rate of the plurality of arithmetic units used for executing the program Provide execution method.

  According to the present invention, in an endoscope apparatus including a plurality of arithmetic units capable of parallel processing, there is an effect that program processing can be appropriately distributed and executed to the plurality of arithmetic units.

It is a figure which shows the structural example of the endoscope system containing the endoscope apparatus which concerns on 1st Embodiment. It is a figure which shows the internal structural example of the image process part and system control part which concern on 1st Embodiment. 4 is a flowchart illustrating an example of operations of an image processing unit and a system control unit according to the first embodiment. It is a figure which shows the internal structural example of the image process part and system control part which concern on 2nd Embodiment. It is a flowchart which shows an example of operation | movement of the image process part and system control part which concern on 2nd Embodiment.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
<First Embodiment>
FIG. 1 is a diagram illustrating a configuration example of an endoscope system including an endoscope apparatus according to the first embodiment of the present invention. This endoscope system is used in, for example, medical institutions (hospitals and the like).

As shown in FIG. 1, the endoscope system 1 includes an endoscope 2, a cable 3, and an endoscope apparatus 4, and the endoscope 2 is connected to the endoscope apparatus 4 via the cable 3. Has a configuration.
The endoscope 2 includes a scope control unit 21, an illumination optical system 22, an objective optical system 23, and an image sensor 24. The scope control unit 21 controls the imaging element 24 according to an imaging control signal input via the cable 3 from a system control unit 412 of a video processor function unit 41 (to be described later) of the endoscope apparatus 4. The illumination optical system 22 illuminates a subject with illumination light emitted from a light source 422 of a light source function unit 42 (to be described later) of the endoscope apparatus 4 and guided through a light guide (not shown). The objective optical system 23 forms an image of reflected light from the subject (reflected light of illumination light illuminated on the subject) on the imaging surface of the imaging device 24. Under the control of the scope control unit 21, the image sensor 24 converts a subject image (reflected light from the subject) formed on the imaging surface into a video signal that is an electrical signal, and the video signal is converted to the cable 3. To the image processing unit 411 of the video processor function unit 41 (to be described later) of the endoscope apparatus 4.

The cable 3 includes a signal line for transmitting a signal (video signal, imaging control signal) transmitted / received between the endoscope 2 and the endoscope apparatus 4.
The endoscope apparatus 4 includes a video processor function unit 41, a light source function unit 42, a monitor 43, a recording medium 44, and an operation unit 45.

  The video processor function unit 41 includes an image processing unit 411 and a system control unit 412. The image processing unit 411 is a GPU having a CKE function, and processes a video signal input from the image sensor 24 via the cable 3 under the control of the system control unit 412. The system control unit 412 controls each unit of the endoscope system 1 such as the image processing unit 411, the light source control unit 421 of the light source function unit 42, and the scope control unit 21 of the endoscope 2. In addition, the system control unit 412 performs control according to the user operation received by the operation unit 45.

  The light source function unit 42 includes a light source control unit 421 and a light source 422. The light source control unit 421 outputs a light source control signal to the light source 422 under the control of the system control unit 412 to control the light source 422. The light source 422 emits illumination light according to the light source control signal input from the light source control unit 421.

  The monitor 43 displays a video corresponding to the video signal processed by the image processing unit 411. The recording medium 44 records a video corresponding to the video signal processed by the image processing unit 411. The operation unit 45 receives various operations from the user and notifies the system control unit 412.

  In the endoscope system 1, the video processor function unit 41 and the light source function unit 42 may be configured integrally as illustrated in FIG. 1 or may be configured separately. In the latter case, the light source function unit 42 may be provided as a separate device with respect to the endoscope device 4. Or you may comprise the endoscope 2, the cable 3, and the endoscope apparatus 4 as integral.

The endoscope 2 may be a flexible endoscope, a rigid endoscope, or a camera head connected to an optical endoscope (fiber scope or surgical optical tube).
The image sensor 24 may be an image sensor such as a charge-coupled device (CCD) or a complementary metal oxide semiconductor (CMOS).

  Further, the illumination method for illuminating the subject may be a simultaneous method, a frame sequential method, or the like. Here, the simultaneous method is a method of illuminating a subject with white light, and the frame sequential method is a method of sequentially illuminating light having different wavelength bands in time series.

  The light source 422 may be a semiconductor light source or a lamp light source. Here, the semiconductor light source is, for example, a white LED that emits white light, a light source that uses a different semiconductor light source for each color component of RGB (red, green, and blue) (a light source that has R, G, and B LEDs), or A laser light source or the like. Alternatively, when the endoscope 2, the cable 3, and the endoscope apparatus 4 are integrally configured as described above, a semiconductor light source as the light source 422 is configured to be provided at the distal end portion of the endoscope 2. May be.

Further, transmission / reception of signals between the endoscope 2 and the endoscope apparatus 4 may be performed by wire as shown in FIG. 1, or configured to be performed wirelessly. May be.
The monitor 43 may be a liquid crystal monitor or the like. Further, the monitor 43 may be configured to be built in the endoscope apparatus 4 as illustrated in FIG. 1 or may be configured to be externally attached to the endoscope apparatus 4.

  Further, the recording medium 44 may be configured to be built in the endoscope apparatus 4 as shown in FIG. 1, or the recording medium 44 may be detachably attached to the endoscope apparatus 4 from an SD card or USB ( Universal Serial Bus) It may be configured as a recording medium such as a memory. Alternatively, the recording medium 44 is used as a recording medium of an external recording apparatus locally connected to the endoscope apparatus 4 or a recording medium of an external recording apparatus (server or the like) connected to the endoscope apparatus 4 via a network. It may be configured.

  Further, the operation unit 45 may be configured to be built in the endoscope apparatus 4 as illustrated in FIG. 1 or may be configured to be externally attached to the endoscope apparatus 4. In the former case, the operation unit 45 may be configured as a front panel switch or the like provided on the front surface of the endoscope apparatus 4. In this case, the front panel switch may be a physical switch or a virtual switch realized by a touch panel or the like. Further, when the operation unit 45 is configured to be externally attached to the endoscope apparatus 4, the operation unit 45 includes a keyboard connected to the endoscope apparatus 4 and a scope provided in the endoscope 2. You may comprise as a switch or the portable information terminal device (Tablet type PC (Personal Computer), a smart phone, etc.) etc. which are connected to the endoscope apparatus 4 by wire or radio | wireless.

  FIG. 2 is a diagram illustrating an internal configuration example of the image processing unit 411 and the system control unit 412. However, here, as an example of the internal configuration of the system control unit 412, only elements mainly related to the image processing unit 411 (elements mainly related to processing of video signals) are shown.

  As shown in FIG. 2, the image processing unit 411 includes a plurality of processing units 4111 for processing a video signal. Each processing unit 4111 includes a plurality of VALUs and executes kernel processing. Hereinafter, the plurality of processing units 4111 are referred to as “processing unit group”.

  Further, the image processing unit 411 has a function of executing the kernel instructed to be executed by the system control unit 412 using the processing unit 4111 that is not used at that time. Further, the image processing unit 411 has a function of calculating a usage rate (usage amount) of the processing unit group and notifying the system control unit 412 of the usage rate. Further, when the kernel is executed, the image processing unit 411 has a function of calculating the usage rate (usage amount) of the processing unit group used for executing the kernel and notifying the usage rate to the system control unit 412. . The usage rate of the processing unit group used for executing the kernel is the same (or substantially the same) usage rate if the kernels are the same.

  The system control unit 412 includes a first usage rate acquisition unit 4121, a second usage rate acquisition unit 4122, a usage rate storage unit 4123, a memory 4124, a kernel selection unit 4125, and a kernel execution instruction unit 4126.

  The first usage rate acquisition unit 4121 acquires the usage rate of the processing unit group of the image processing unit 411 from the image processing unit 411. When the kernel is executed by the image processing unit 411, the second usage rate acquisition unit 4122 acquires the usage rate of the processing unit group used to execute the kernel from the image processing unit 411.

  The usage rate storage unit 4123 stores the usage rate of the processing unit group used for executing the kernel for each kernel executed by the image processing unit 411 in the past. For example, although not shown, in the usage rate storage unit 4123, the usage rate of the processing unit group used for the execution of the past color correction processing kernel is 30%, and used for the execution of the past brightness correction processing kernel. The processing unit group usage rate is 20%, the past image composition processing kernel usage rate is 10%, and the past JPEG (Joint Photographic Experts Group) compression processing kernel is used. The usage rate of the processing unit group for each kernel executed in the past is stored such that the usage rate of the processing unit group is 10%. Note that the usage rate of the processing unit group for each kernel is acquired by the second usage rate acquisition unit 4122.

  The memory 4124 stores (stores) a kernel waiting for execution. In the example shown in FIG. 2, “kernel # 1” (color correction processing kernel), “kernel # 2” (brightness correction processing kernel), “kernel # 3” (image composition processing kernel) in the order of storage in the memory 4124. ) And “kernel # 4” (JPEG compression processing kernel) are stored. Note that the storage order of kernels in the memory 4124 is also the order of kernels waiting to be executed.

  The kernel selection unit 4125 is based on the usage rate of the processing unit group acquired by the first usage rate acquisition unit 4121 and the usage rate of the processing unit group for each kernel stored in the usage rate storage unit 4123. A kernel to be executed next is selected (determined) from the execution waiting kernels stored in the memory 4124. For example, the kernel selection unit 4125 selects a kernel that can be immediately started by the image processing unit 411 and that is stored first in the memory 4124 as a kernel to be executed next. . In this selection, the usage rate of the processing unit group used for executing the kernel is the same (or substantially the same) usage rate if the kernels are the same. The usage rate of the processing unit group used to execute each of the kernels can be predicted from the usage rate of the processing unit group for the same kernel stored in the usage rate storage unit 4123. Further, the unused rate of the processing unit group at that time can be obtained from the usage rate of the processing unit group acquired by the first usage rate acquiring unit 4121. Therefore, a kernel that can be immediately started to be executed by the image processing unit 411 among the waiting kernels can be a kernel in which the usage rate of the predicted processing unit group falls below the unused rate of the processing unit group. .

  For example, the usage rate of the processing unit group acquired by the first usage rate acquisition unit 4121 is 80% (that is, the unused rate is 20%), and the processing unit for each kernel stored in the usage rate storage unit 4123 It is assumed that the usage rate of the group is exemplified above, and the execution waiting kernel stored in the memory 4124 is exemplified in FIG. In this case, among the kernels waiting to be executed, the kernels that can be immediately started by the image processing unit 411 are “kernel # 2” (brightness correction processing), “kernel # 3” (image synthesis processing kernel), and “kernel # 2”. Kernel # 4 "(JPEG compression kernel). In this case, “kernel # 2” (brightness correction processing) that is the first storage order in the memory 4124 is selected by the kernel selection unit 4125 as the kernel to be executed next. Will be.

  Alternatively, for example, the usage rate of the processing unit group acquired by the first usage rate acquisition unit 4121 is 80% (that is, the unused rate is 20%), and for each kernel stored in the usage rate storage unit 4123 The usage rate of the processing unit group is exemplified above, and the execution waiting kernels stored in the memory 4124 are “kernel # 1” (color correction processing kernel), “kernel # 2” in the order of storage in the memory 4124. "(Image composition processing kernel)" and "kernel # 3" (JPEG compression processing kernel). In this case, among the kernels waiting to be executed, the kernels that can be immediately started by the image processing unit 411 are “kernel # 2” (image composition processing kernel) and “kernel # 3” (JPEG compression processing kernel). . In this case, the kernel selection unit 4125 selects “kernel # 2” (image composition processing kernel), which is the first kernel stored in the memory 4124, as the kernel to be executed next. Will be. In this case, the unused rate of the processing unit group is 20%, and the usage rate of each processing unit group of the image synthesis processing kernel and the JPEG compression processing kernel is predicted to be 10%. The unit 4125 may select both “kernel # 2” (image composition processing kernel) and “kernel # 3” (JPEG compression processing kernel) as kernels to be executed next.

Then, the kernel selection unit 4125 transfers the kernel selected in this way to the kernel execution instruction unit 4126.
The kernel execution instruction unit 4126 instructs the image processing unit 411 to execute the kernel transferred by the kernel selection unit 4125.

  In the endoscope apparatus 4 included in the endoscope system 1 configured as described above, the processing unit group (a plurality of processing units 4111) is an example of a plurality of arithmetic units. The first usage rate acquisition unit 4121 is an example of a first acquisition unit that acquires usage rates of a plurality of arithmetic units. The second usage rate acquisition unit 4122 is an example of a second acquisition unit that acquires usage rates of a plurality of arithmetic units used to execute the program when the program is executed. The usage rate storage unit 4123 is an example of a storage unit that stores usage rates of a plurality of arithmetic units used for executing the program for each program executed in the past. Based on the usage rate acquired by the first acquisition unit and the usage rate for each program stored in the storage unit, the kernel selection unit 4125 selects a program to be executed next from among the programs waiting to be executed. It is an example of the determination part to determine.

  Next, operations of the image processing unit 411 and the system control unit 412 illustrated in FIG. 2 will be described as an example of the operation of the endoscope system 1. In the description of this operation, the image processing unit 411 is described as the GPU 411.

FIG. 3 is a flowchart showing an example of the operation.
As shown in FIG. 3, in this operation, first, the GPU 411 calculates the usage rate of the processing unit group (S100).

Subsequently, the first usage rate acquisition unit 4121 acquires the usage rate calculated in S100 from the GPU 411 (S105), and notifies the kernel selection unit 4125 (S110).
In S100 and S105, for example, the first usage rate acquisition unit 4121 inquires of the GPU 411 about the usage rate, the GPU 411 performs the processing of S100 in response to the inquiry, and acquires the first usage rate. The unit 4121 may perform the process of S105.

  Subsequently, the kernel selection unit 4125 is used to execute each kernel waiting for execution stored in the memory 4124 based on the usage rate of the processing unit group for each kernel stored in the usage rate storage unit 4123. The usage rate of the processing unit group is predicted (S115).

Subsequently, the kernel selection unit 4125 determines whether or not the prediction has been completed (S120).
When the determination result in S120 is No, the kernel selection unit 4125 selects one kernel waiting for execution for which the usage rate of the processing unit group could not be predicted, and transfers the selected kernel to the kernel execution instruction unit 4126. Is executed by the kernel execution instructing unit 4126 to the GPU 411 (S125). Subsequently, the GPU 411 executes the kernel designated by the kernel execution instruction unit 4126, and calculates the usage rate of the processing unit group used for executing the kernel (S130). Subsequently, the second usage rate acquisition unit 4122 acquires the usage rate calculated in S130 from the GPU 411, and the usage rate is stored as the usage rate of the processing unit group of the kernel instructed to be executed in S125. Stored in the unit 4123 (S135), and the process returns to S115.

  On the other hand, when the determination result in S120 is Yes, the kernel selection unit 4125 waits for execution stored in the memory 4124 based on the usage rate of the processing unit group notified in S110 and the prediction performed in S115. Among the kernels, the kernel in which the GPU 411 can start execution immediately is selected, and the kernel having the earliest storage order in the memory 4124 is selected as the kernel to be executed next (S140). However, if there is no kernel that can be immediately executed by the GPU 411 among the kernels waiting for execution stored in the memory 4124, the kernel having the earliest storage order in the memory 4124 is selected.

Subsequently, the kernel selection unit 4125 transfers the selected kernel to the kernel execution instruction unit 4126 (S145).
Subsequently, the kernel execution instruction unit 4126 instructs the GPU 411 to execute the kernel transferred by the kernel selection unit 4125 in S145 (S150).

  Subsequently, the GPU 411 executes the kernel instructed by the kernel execution instructing unit 4126 in S150 using the unused processing unit 4111 (S155), and this flow ends.

  As described above, in the endoscope apparatus 4 according to this embodiment, the image processing unit (GPU) 411 is a kernel that can immediately start execution from the execution waiting kernels stored in the memory 4124. Since the kernel having the earliest storage order in the memory 4124 is selected as the kernel to be executed next, the image processing unit (GPU) 411 uses kernel processing. The processing units 4111 that are not present are distributed without delay and are immediately executed. Therefore, according to the endoscope apparatus 4 according to the present embodiment, kernel processing can be appropriately distributed and executed among processing unit groups (a plurality of processing units 4111). As a result, the video signal can be processed at a higher speed.

<Second Embodiment>
The endoscope apparatus according to the second embodiment of the present invention must be executed within the processing time limit in the kernel waiting for execution in the endoscope apparatus 4 according to the first embodiment. If there is a kernel, it is executed with priority.

In the description of the second embodiment, the same elements as those described in the first embodiment are denoted by the same reference numerals, and detailed description thereof is omitted.
FIG. 4 is a diagram illustrating an internal configuration example of the image processing unit (GPU) 411 and the system control unit 412 in the endoscope system 1 according to the second embodiment. FIG. 4 is also a diagram corresponding to the internal configuration example shown in FIG.

  4 is different from the first embodiment (internal configuration example shown in FIG. 2) in that the system control unit 412 further includes a processing time limit setting unit 4127 for setting a processing time limit. Note that this processing time limit is, for example, the maximum time allowed for processing for one frame. For example, when acquiring 60 frames / second video, the processing time limit is set to 1/60 seconds. Further, the processing time limit set by the processing time limit setting unit 4127 can be changed, for example, by an operation of the operation unit 45 by the user.

  Along with the provision of this processing time limit setting unit 4127, in the second embodiment, the kernel selection unit 4125 further includes a processing time limit setting unit 4127 in the execution waiting kernel stored in the memory 4124. If there is a kernel that must be executed within the set processing time limit, the kernel having the earliest storage order in the memory 4124 is selected with priority as the kernel to be executed next ( Determined) and transferred to the kernel execution instruction unit 4126. Here, the kernel that must be executed within the processing time limit set by the processing time limit setting unit 4127 is, for example, a kernel that must be executed in the processing for one frame, and is JPEG compression processing. Kernel etc. are excluded.

About another structure, it is the same as that of 1st Embodiment.
Next, operations of the image processing unit (GPU) 411 and the system control unit 412 illustrated in FIG. 4 will be described as an example of the operation of the endoscope system 1.

FIG. 5 is a flowchart showing an example of the operation. FIG. 5 is also a diagram corresponding to the flowchart shown in FIG.
In FIG. 5, the difference from the first embodiment (the flowchart shown in FIG. 3) is that when the determination result of S120 is Yes, the determination of S200 is performed, and when the determination result of S200 is No, the process proceeds to S140. In the case of Yes and Yes, the process of S205 is performed and then the process proceeds to S145.

  More specifically, when the determination result in S120 is Yes, the kernel selection unit 4125 within the processing time limit set by the processing time limit setting unit 4127 in the execution waiting kernel stored in the memory 4124. It is determined whether there is a kernel to be executed (S200). Here, when the determination result is No, the process proceeds to S140.

  On the other hand, when the determination result in S200 is Yes, the kernel selection unit 4125 executes the processing within the processing time limit set by the processing time limit setting unit 4127 from the execution waiting kernels stored in the memory 4124. The kernel that must be stored and that has the earliest storage order in the memory 4124 is selected as the kernel to be executed next (S205), and the process proceeds to S145.

About others, it is the same as that of 1st Embodiment.
As described above, according to the endoscope device 4 according to the second embodiment, in addition to the effects described in the first embodiment, the execution is performed within the processing time limit in the kernel waiting for execution. If there is a kernel that must be present, the image processing unit (GPU) 411 can be made to execute it preferentially.

In the above-described embodiments, the following modifications are possible.
For example, when the first usage rate acquisition unit 4121 acquires the usage rate of the processing unit group from the image processing unit (GPU) 411, which processing unit 4111 is in use (or unused) is determined. Information (hereinafter referred to as “processing unit group use state information”) may be acquired from the image processing unit (GPU) 411. Then, when the kernel execution instructing unit 4126 instructs the image processing unit (GPU) 411 to execute the kernel, the processing unit 4111 ( The image processing unit (GPU) 411 may be instructed to execute using the unused processing unit 4111).

  For example, the video processor function unit 41 may include a plurality of image processing units (GPUs) 411. In this case, the system control unit 412 controls the image processing unit (GPU) 411 described in each of the above-described embodiments for each of the plurality of image processing units (GPU) 411.

  As mentioned above, the embodiment described above shows a specific example of the present invention in order to facilitate understanding of the invention, and the present invention is not limited to the above-described embodiment. The present invention can be variously modified and changed without departing from the concept of the present invention defined in the claims.

DESCRIPTION OF SYMBOLS 1 Endoscope system 2 Endoscope 3 Cable 4 Endoscope apparatus 21 Scope control part 22 Illumination optical system 23 Objective optical system 24 Image sensor 41 Video processor function part 42 Light source function part 43 Monitor 44 Recording medium 45 Operation part 411 Image Processing unit 412 System control unit 421 Light source control unit 422 Light source 4111 Processing unit 4121 First usage rate acquisition unit 4122 Second usage rate acquisition unit 4123 Usage rate storage unit 4124 Memory 4125 Kernel selection unit 4126 Kernel execution instruction unit 4127 Processing restriction Time setting section

Claims (6)

Multiple arithmetic units;
A first acquisition unit for acquiring a usage rate of the plurality of arithmetic units;
A storage unit that stores a usage rate of each of the plurality of arithmetic units used for executing the program for each program executed in the past;
A determination unit that determines a program to be executed next from among the programs waiting to be executed based on the usage rate acquired by the first acquisition unit and the usage rate for each program stored in the storage unit When,
An endoscope apparatus comprising: When the program is executed, the system further comprises a second acquisition unit that acquires usage rates of the plurality of arithmetic units used for executing the program,
The usage rate acquired by the second acquisition unit is stored in the storage unit.
The endoscope apparatus according to claim 1. The determining unit determines a program that can immediately start execution and that has the earliest waiting order for execution as a program to be executed next.
The endoscope apparatus according to claim 1 or 2, characterized by the above-mentioned. The determination unit, when there is a program that must be executed within a predetermined time among the programs waiting for execution, determines the program as a program to be executed next with priority.
The endoscope apparatus according to any one of claims 1 to 3, wherein the endoscope apparatus is characterized in that The plurality of arithmetic units are a plurality of arithmetic units included in a GPU (Graphics Processing Unit).
The endoscope apparatus according to any one of claims 1 to 4, wherein the endoscope apparatus is characterized in that: An endoscope apparatus program execution method including a plurality of arithmetic units,
Obtaining a usage rate of the plurality of arithmetic units;
Based on the acquired usage rate of the plurality of arithmetic units and the usage rate of the plurality of arithmetic units used to execute the program for each program executed in the past, from among the programs waiting to be executed Determine the next program to run,
A program execution method characterized by the above.

Images