KR20230053278A - Raw data purification method for artificial intelligence learning according to importance evaluation and a computer program recorded on a recording medium to execute the same - Google Patents

Abstract

The present invention provides a data cleansing method of a device for collecting data for artificial intelligence (AI) machine learning. The method of the present invention comprises the steps of: allowing a learning data collection device to use a sensor to collect data; allowing the learning data collection device to cleanse the collected data; and allowing the learning data collection device to evaluate the importance of the collected data; and allowing the learning data collection device to transmit the cleansed data to the data collection server. According to the present invention, the cleansing step can be performed by calculating the similarity between an object included in the data and an object subject to AI learning.

Description

Raw data purification method for artificial intelligence learning according to importance evaluation and a computer program recorded on a recording medium to execute the raw data purification method for artificial intelligence learning according to importance evaluation same}

The present invention relates to the collection of data for artificial intelligence (AI) machine learning. More specifically, it relates to a method for refining raw data for artificial intelligence learning according to the importance evaluation of data for artificial intelligence (AI) machine learning and a computer program recorded on a recording medium to execute the method.

Artificial intelligence (AI) refers to a technology that artificially implements some or all of human learning abilities, reasoning abilities, and perception abilities using computer programs. In relation to artificial intelligence (AI), machine learning refers to learning to optimize parameters with given data using a model composed of multiple parameters. Such machine learning is classified into supervised learning, unsupervised learning, and reinforcement learning according to the form of learning data.

In general, designing data for artificial intelligence (AI) learning proceeds in the steps of data structure design, data collection, data refinement, data processing, data expansion, and data verification.

To describe each step in more detail, data structure design is performed through ontology definition, classification system definition, and the like. Data collection is performed by collecting data through direct filming, web crawling, or associations/professional organizations. Data purification is performed by removing redundant data from collected data and de-identifying personal information. Data processing is performed by performing annotation and inputting metadata. Data extension is performed by performing ontology mapping and supplementing or extending the ontology as needed. In addition, data verification is performed by verifying validity according to the set target quality using various verification tools.

On the other hand, autonomous driving of a vehicle refers to a system that can judge and drive a vehicle by itself. Such autonomous driving may be classified into gradual stages from non-automation to complete automation according to the degree of involvement of the system in driving and the degree of control of the vehicle by the driver. In general, the level of autonomous driving is divided into six levels classified by the Society of Automotive Engineers (SAE) International. According to the six levels classified by the International Society of Automotive Engineers (SAE), level 0 is no automation, level 1 is driver assistance, level 2 is partial automation, and level 3 The stage is conditional automation, level 4 is high automation, and level 5 is full automation.

Autonomous driving of vehicles is performed through mechanisms of perception, localization, path planning, and control. Currently, several companies are developing to implement recognition and path planning among autonomous driving mechanisms using artificial intelligence (AI).

Data used for machine learning of artificial intelligence (AI) that can be used for autonomous driving of these vehicles is collected by various types of sensors installed in the vehicle. For example, the data used for machine learning of artificial intelligence (AI) that can be used for autonomous driving of a vehicle is lidar, camera, radar, and ultrasonic sensor fixed to the vehicle. ) and GPS (Global Positioning System), etc., may be acquired, photographed, or sensed data, but are not limited thereto.

However, when data is collected in real time using various types of sensors installed in vehicles for use in artificial intelligence (AI) machine learning, the volume of the collected data increases exponentially, but the bandwidth of mobile communication and There is a difficulty in transmitting data collected by region and the like to the outside in real time.

In addition, data collected by various sensors may include useless data or personal information. Furthermore, when the physical conditions of a vehicle equipped with various sensors for collecting data and a vehicle to perform autonomous driving are different, the performance of autonomous driving by machine learning artificial intelligence (AI) can be improved using the collected data. There are limitations that cannot be guaranteed.

The data collection server performs de-identification and redundant data removal on numerous raw data and provides them to the annotation server.

However, the importance of raw data processed by the data collection server is all different. For example, images containing objects directly related to artificial intelligence learning may have high importance, but images without any objects may have low importance.

Therefore, there is a need to improve the efficiency of data by refining raw data of relatively low importance according to the AI learning goal.

Korean Patent Laid-open Publication No. 10-2020-0042629, ‘Method and apparatus for generating touch-based annotations and images in mobile devices for artificial intelligence learning’, (published on April 24, 2020)

One object of the present invention is to provide a raw data purification method for artificial intelligence learning according to the importance evaluation of data for artificial intelligence (AI) machine learning of a vehicle and a computer program recorded on a recording medium to execute the same.

The technical problems of the present invention are not limited to the technical problems mentioned above, and other technical problems not mentioned will be clearly understood by those skilled in the art from the following description.

In order to achieve the technical problem as described above, the present invention proposes a data purification method of a data collection device for machine learning (Artificial Intelligence, AI). The method includes the steps of, by a learning data collection device, collecting data using a sensor; refining the collected data by the learning data collection device; and evaluating, by the learning data collection device, importance of the collected data. and transmitting, by the learning data collection device, refined data to a data collection server.

Specifically, the step of refining may be performed by calculating the similarity between the object included in the data and the object to be artificial intelligence learning, and the similarity between the object included in the data and the object to be artificial intelligence learning It can be refined by removing raw data below the threshold.

In addition, in the refining, the learning data collection device may perform purification by removing raw data that is different from an environment related to artificial intelligence learning based on information about a capturing time point of the raw data.

In the present invention, the learning data collection device controls the refinement rate of raw data based on the meta data related to the collection time of the raw data, and the raw data affects artificial intelligence learning based on the collected meta data. impact can be calculated.

In order to achieve the technical problem as described above, the present invention proposes a computer program recorded on a recording medium to execute the purification method as described above. The computer program may include a memory; transceiver; and a processor configured to process instructions residing in the memory, wherein the processor collects data using a sensor; Step, by the processor, refining the collected data; and evaluating, by the processor, importance of the collected data; and transmitting, by the processor, the refined data to a data collection server; in order to execute the, it may be a computer program recorded on a recording medium.

Details of other embodiments are included in the detailed description and drawings.

According to embodiments of the present invention, the learning data collection device can refine raw data of low importance prior to performing de-identification processing, etc., through which data with low efficiency in artificial intelligence learning can be reduced. It works.

In particular, the learning data collection device calculates the similarity between an object included in the raw data and an object to be an artificial intelligence learning target, and removes raw data having a similarity below a threshold to perform purification.

In addition, the learning data collection device may perform purification by removing raw data that is different from the environment related to artificial intelligence learning based on information (GPS, weather information, etc.) on the shooting time of the raw data.

The effects of the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description of the claims.

It should be noted that the technical terms used in this specification are only used to describe specific embodiments and are not intended to limit the present invention. In addition, technical terms used in this specification should be interpreted in terms commonly understood by those of ordinary skill in the art to which the present invention belongs, unless specifically defined otherwise in this specification, and are overly inclusive. It should not be interpreted in a positive sense or in an excessively reduced sense. In addition, when the technical terms used in this specification are incorrect technical terms that do not accurately express the spirit of the present invention, they should be replaced with technical terms that those skilled in the art can correctly understand. In addition, general terms used in the present invention should be interpreted as defined in advance or according to context, and should not be interpreted in an excessively reduced sense.

Also, singular expressions used in this specification include plural expressions unless the context clearly indicates otherwise. In this application, terms such as "consisting of" or "having" should not be construed as necessarily including all of the various components or steps described in the specification, and some of the components or steps are included. It should be construed that it may not be, or may further include additional components or steps.

Also, terms including ordinal numbers such as first and second used in this specification may be used to describe various components, but the components should not be limited by the terms. These terms are only used for the purpose of distinguishing one component from another. For example, a first element may be termed a second element, and similarly, a second element may be termed a first element, without departing from the scope of the present invention.

When a component is referred to as being “connected” or “connected” to another component, it may be directly connected or connected to the other component, but other components may exist in the middle. On the other hand, when a component is referred to as “directly connected” or “directly connected” to another component, it should be understood that no other component exists in the middle.

Hereinafter, preferred embodiments according to the present invention will be described in detail with reference to the accompanying drawings, but the same or similar components are given the same reference numerals regardless of reference numerals, and redundant description thereof will be omitted. In addition, in describing the present invention, if it is determined that a detailed description of a related known technology may obscure the gist of the present invention, the detailed description will be omitted. In addition, it should be noted that the accompanying drawings are only for easily understanding the spirit of the present invention, and should not be construed as limiting the spirit of the present invention by the accompanying drawings. The spirit of the present invention should be construed as extending to all changes, equivalents or substitutes other than the accompanying drawings.

The data collected for artificial intelligence learning consists of very large numbers taken in real time, not just a few. There is a lot of burden for the learning data collection device to transmit a lot of raw data to the data collection server for the collected data. In addition, since the collection device generally collects raw data through real-time shooting, among the collected raw data, similar data with low artificial intelligence learning efficiency may be included, thereby highlighting the efficiency of data management.

An artificial intelligence learning system according to an embodiment of the present invention may include a plurality of learning data collection devices, a learning data generating device, a plurality of annotation devices, and an artificial intelligence learning device.

Since the components of the artificial intelligence learning system according to an embodiment are merely functionally distinct elements, two or more components are integrated and implemented in an actual physical environment, or one component is implemented in an actual physical environment. may be implemented separately from each other.

The learning data collection device according to the present invention is described based on being applied to a vehicle, but is not limited to being installed in a vehicle, and is a concept including various embodiments for collecting artificial intelligence learning data.

To describe each component, the learning data collection device collects data for machine learning of artificial intelligence (AI) that can be used for autonomous driving. A device that collects data in real time from one or more of a radar and an ultrasonic sensor.

Characteristically, the learning data collection device according to various embodiments of the present invention can efficiently control a plurality of cameras installed at different heights in a vehicle and simultaneously taking pictures in the same direction in real time. The learning data collecting device may automatically detect objects included in the collected data and automatically de-identify personal information before transmitting the collected data to the learning data generating device. In addition, the learning data collecting device can efficiently transmit the collected data according to communication quality with the learning data generating device.

Types of sensors that are controlled by the learning data collection device and are installed in the vehicle to acquire, photograph, or detect machine learning data include lidar, camera, radar, and ultrasonic sensor. ), but may include one or more of them, but is not limited thereto. In addition, sensors that are controlled by the learning data collection device and are installed in the vehicle to acquire, photograph, or detect machine learning data are not limited to being provided one by one for each type, and may be provided in plural even if they are of the same type. .

With the following configuration, the learning data generating device receives data collected in real time by each learning data collecting device from each of the plurality of learning data collecting devices using mobile communication, and converts the received data into a plurality of annotations. It is a device that distributes to devices and generates data for machine learning of artificial intelligence (AI) based on the annotation work result received from each annotation device.

Specifically, the learning data generating device may receive attributes of a project related to artificial intelligence (AI) learning from the artificial intelligence learning device. The learning data generation device performs data structure design, collected data refinement, data processing, data expansion, and data verification for artificial intelligence (AI) learning, based on the user's control and the attributes of the received project. can do.

First of all, the learning data generating device may design a data structure for artificial intelligence (AI) learning. For example, the learning data generating device may define an ontology for artificial intelligence (AI) learning and a classification system of data for artificial intelligence (AI) learning based on the properties of the user's control and the received project. can

The learning data generating device may collect data for artificial intelligence (AI) learning based on the designed data structure. To this end, the learning data generating device may receive sensing data, 3D point cloud data, 2D images, and distance information from the learning data collecting device. However, it is not limited thereto, and the learning data generating device may perform web crawling or download data from an external organization's device.

The learning data generating apparatus may remove redundant or extremely similar data from among the collected sensing data, 3D point cloud data, 2D images, and distance information. The learning data generating device may verify whether the personal information automatically de-identified for the 2D image by the learning data collecting device is correct.

On the other hand, in order to support the automation of object detection performed by the learning data collection device, the learning data generating device uses object detection artificial intelligence (AI) machine-learned based on the result of an existing annotation task targeting a plurality of 2D images. can be provided In addition, the learning data generating device may generate an object detection rule using object detection artificial intelligence (AI) and distribute the generated object detection rule to a plurality of learning data collection devices.

In this case, the object detection rule may be a rule enumerating edge patterns classified by object type so as to verify the object identified in the 2D image.

The learning data generating device may receive an update request for object detection rules from the learning data collecting device. In this case, the update request may include, but is not limited to, a 2D image that has not been properly verified by the existing object detection rule, a type of object identified from the 2D image, and a pattern of an edge extracted for the identified object. The learning data generating device may perform reinforcement learning of object detection artificial intelligence (AI) based on the received update request. Also, when a new object detection rule is generated as a result of reinforcement learning, the learning data generation device may redistribute the new object detection rule to a plurality of learning data collection devices.

Meanwhile, the learning data generating device may distribute and transmit collected and refined sensing data, 3D point cloud data, 2D images, and distance information to a plurality of annotation devices. In this case, the learning data generating device may distribute sensing data, 3D point cloud data, 2D images, and distance information corresponding to a pre-allocated amount for an operator (ie, labeler) of the annotation device.

The learning data generating device may receive annotation work results from each annotation device. The learning data generating device may generate artificial intelligence (AI) learning data by packaging the received annotation work result. And, the learning data generating device may transmit the generated artificial intelligence (AI) learning data to the artificial intelligence learning device.

Any device capable of transmitting and receiving data to and from a learning data collection device, an annotation device, and an artificial intelligence learning device, and performing calculations based on the transmitted and received data may be accepted as the learning data generation device having such characteristics. . For example, the learning data generating device may be any one of a fixed computing device such as a desktop, workstation, or server, but is not limited thereto.

With the following configuration, the annotation device is a device that can be used to perform annotation work on data collected from LIDAR, camera, radar, or ultrasound. That is, the annotation device is a device that can be used to perform annotation work on the sensed data distributed by the learning data generating device, 3D point cloud data, 2D images, and distance information.

All or part of such an annotation device may be a device in which an annotation worker performs annotation work through a clouding service.

Specifically, the annotation device selects one sensed data, 3D point cloud data, 2D image, or distance information to be annotated from among the sensed data, 3D point cloud data, 2D images, and distance information received from the learning data generating device. can be output to the display.

The annotation device may select a tool according to a signal input from a user through an input/output device. Here, the tool is a tool for setting a bounding box specifying one or more objects included in sensing data, 3D point cloud data, 2D image, or distance information.

The annotation device may receive coordinates according to the selected tool through an input/output device. In addition, the annotation device may specify an object included in sensing data, 3D point cloud data, 2D image, or distance information by setting a bounding box based on the input coordinates.

Here, the bounding box is an area for specifying an object to be learned by artificial intelligence (AI) among objects included in sensing data, 3D point cloud data, 2D image, or distance information. Such a bounding box may have a rectangle or cube shape, but is not limited thereto.

The annotation device may generate sensing data, 3D point cloud data, 2D image, distance information, or metadata for a set object to be annotated according to a signal input from a user through an input/output device. Here, the metadata is sensing data, 3D point cloud data, 2D image, distance information, or data for describing an object. Such metadata includes the category of the specified object, the rate at which the object is clipped by the angle of view, the rate at which the object is obscured by other objects or objects, the tracking ID of the object, the time the image was taken, and the weather conditions on the day the image was taken. may include, but are not limited to, file size, image size, copyright holder, resolution, bit value, aperture transmittance, exposure time, ISO sensitivity, focal length, aperture value, angle of view, white balance, RGB depth, class name , tag, shooting location, type of road, road surface information, or traffic jam information may be further included.

The annotation device may generate an annotation work result based on the specified object and generated metadata. In this case, the annotation work result may have a JSON (Java Script Object Notation) file format, but is not limited thereto. The annotation device may transmit the generated annotation work result to the learning data generating device.

As the annotation device having such characteristics, any device may be accepted as long as it is capable of transmitting and receiving data to and from the learning data generating device and performing calculations based on the transmitted and received data. For example, the annotation device may be a stationary computing device such as a desktop, workstation, or server, or a smartphone, laptop, tablet, or phablet. ), a portable multimedia player (PMP), a personal digital assistant (PDA), or an e-book reader.

With the following configuration, the artificial intelligence learning device is a device that can be used for machine learning of artificial intelligence (AI) that can be used for autonomous driving of a vehicle.

Specifically, the artificial intelligence learning device may transmit requirements for achieving the purpose of artificial intelligence (AI) that can be used for autonomous vehicle driving to the learning data generating device. The artificial intelligence learning device may receive artificial intelligence (AI) learning data from the learning data generating device. In addition, the artificial intelligence learning device may perform machine learning of artificial intelligence (AI) that can be used for autonomous driving of a vehicle using the received artificial intelligence (AI) learning data.

As such, the artificial intelligence learning device may be any device capable of transmitting and receiving data to and from the learning data generating device and performing calculations using the transmitted and received data. For example, the artificial intelligence learning device may be any one of a desktop, a workstation, or a fixed computing device such as a server, but is not limited thereto.

As described above, the plurality of learning data collection devices and learning data generating devices may transmit and receive data using a mobile communication network. For example, in a mobile communication network, Code Division Multiple Access (CDMA), Wideband CDMA (WCDMA), High Speed Packet Access (HSPA), Long Term Evolution (Long Term Evolution) Evolution, LTE) and 5th generation mobile telecommunication may be included, but is not limited thereto.

In addition, the learning data generating device, the plurality of annotation devices, and the artificial intelligence learning device may transmit and receive data using a network in which one or more of a security line, a common wired communication network, or a mobile communication network directly connects devices. For example, public wired communication networks may include Ethernet, x Digital Subscriber Line (xDSL), Hybrid Fiber Coax (HFC), and Fiber To The Home (FTTH). It may be, but is not limited thereto.

Hereinafter, sensors that are controlled by the above-described learning data collection device and are installed in a vehicle to acquire, photograph, or sense machine learning data will be described in more detail.

An apparatus for collecting learning data according to an embodiment of the present invention may include a communication unit, an input/output unit, a multi-sensor control unit, an object identification unit, a de-identification processing unit, a data providing unit, and a storage unit.

Since the components of the learning data collection device are merely functionally distinct elements, two or more components are integrated and implemented in an actual physical environment, or one component is implemented separately from each other in an actual physical environment. It could be.

Describing each component, the communication unit may transmit/receive data between multiple sensors installed in the vehicle and the learning data generating device.

Specifically, the communication unit may receive sensing data, 3D point cloud data, 2D image, and distance information from a radar, lidar, camera, and ultrasonic sensor fixedly installed in a vehicle.

The communication unit may measure communication quality of mobile communication with the learning data generating device at predetermined intervals. For example, the communication unit may transmit an Internet Control Message Protocol (ICMP) ping packet (echo request) to the learning data generating device at predetermined intervals and receive a response packet (echo reply) from the learning data generating device. .

The communication unit may transmit sensing data, 3D point cloud data, 2D image, and distance information to the learning data generating device under the control of the data providing unit. In this case, the 2D image may include information about a region occupied by an object identified from the 2D image by the object identification unit and information about de-identified personal information.

The communication unit may transmit an update request message of an object detection rule or an update request message of a de-identification processing rule to the learning data generating device under the control of the object identification unit and the de-identification processing unit.

With the following configuration, the input/output unit may receive a signal from a user through a user interface (UI) or output an operation result to the outside.

Specifically, the input/output unit may receive a period for measuring communication quality with the learning data generating device from the user. The input/output unit may receive input from a user of a basic size of a buffer for storing 3D point cloud data, 2D image, sensing data, and distance information, and a priority of data to be preferentially stored in the buffer.

With the following configuration, the multi-sensor controller may control one or more of a radar, lidar, camera, and ultrasonic sensor fixedly installed in a vehicle, and collect data from one or more of the radar, lidar, camera, and ultrasonic sensor in real time.

And, the multi-sensor controller may store data collected in real time from at least one of radar, lidar, camera, and ultrasonic sensor in a buffer of the data providing unit.

Basically, the data providing unit may measure the communication quality of mobile communication with the learning data generating device at predetermined intervals and adjust the size of the buffer in response to the measured communication quality.

Specifically, the data providing unit transmits ICMP ping packets to the learning data generating device at set intervals through the communication unit, and RTT (Round Trip Time) and TTL (Time To Live) included in the response packet received from the learning data generating device ), the size of the buffer can be determined based on In this case, the data providing unit may adjust a period for transmitting ICMP ping packets in inverse proportion to a moving speed of a vehicle equipped with a radar, lidar, camera, and ultrasonic sensor.

On the other hand, when it is determined that all packets are lost based on the TTL included in the response packet, the data providing unit does not directly perform mobile communication with the learning data generating device, but communicates with the learning data generating device via another learning data collecting device. It can also be controlled to perform mobile communication.

If the size of the buffer determined by the data provider is smaller than the preset basic transmission size, the data provider may discard some of the data collected from one or more of radar, lidar, camera, and ultrasonic sensor. there is. In this case, the basic transmission size may be the size of data obtained by summing all data simultaneously acquired, photographed, and sensed by radar, LIDAR, camera, and ultrasonic sensor during a preset basic collection time.

First of all, the data providing unit sequentially transmits detection data acquired by radar, 3D point cloud data acquired by lidar, 2D image captured by camera, and distance information detected by ultrasonic sensor according to preset priorities. can be removed

In this case, the data provider estimates the distance the vehicle is separated from the object based on the detection data obtained by the radar, and based on the estimated distance, the detection data obtained by the radar and the 3D point cloud data obtained by the LIDAR , it is possible to adjust the priority between the 2D image captured by the camera and the distance information sensed by the ultrasonic sensor.

In addition, when the data providing unit needs to remove 2D images captured by a camera according to priority, a 2D image that does not include an object may be preferentially removed from among a plurality of 2D images captured by a camera during the basic collection time. there is.

With the following configuration, the storage unit may store data necessary for the operation of the learning data collection device.

Specifically, the storage unit may include a buffer for storing 3D point cloud data, 2D image, sensing data, and distance information. The storage unit may include a database for storing rules and basic data required for image processing and data transmission.

Object detection rules, non-identification processing rules, and 3D models for each type of object may be stored in the database constituting the storage unit. Here, the object detection rule may be a rule listing edge patterns classified according to object types so as to verify objects identified in the 2D image. The de-identification processing rule may be a rule enumerating edge patterns of de-identification areas existing in an object so as to determine a de-identification process area in a 2D image. Also, the 3D model is three-dimensional shape data of objects classified according to object types.

Hereinafter, hardware for implementing the above-described logical components of the learning data collection device will be described in more detail.

The learning data collection device may include a processor, memory, transceiver, input/output device, data bus, and storage.

The processor may implement operations and functions of the learning data collection device based on instructions according to software in which the method according to the embodiments of the present invention is implemented, which resides in the memory. Software in which a method according to embodiments of the present invention is implemented may be loaded in the memory. The transceiver may transmit and receive data to and from radar, lidar, cameras, ultrasonic sensors, and learning data generating devices. The input/output device may receive data necessary for the operation of the learning data collection device and output collected sensing data, 3D point cloud data, 2D image, and distance information. The data bus is connected to a processor, a memory, a transceiver input/output device, and a storage, and may serve as a movement path for transferring data between respective components.

The storage may store an application programming interface (API), a library file, a resource file, and the like required for execution of software in which a method according to embodiments of the present invention is implemented. The storage may store software and a database in which a method according to embodiments of the present invention is implemented. Object detection rules, non-identification processing rules, and 3D models for each type of object may be stored in the database.

According to one embodiment of the present invention, software for implementing a method of controlling sensors resident in a memory or stored in a storage includes a processor that collects 2D images simultaneously taken by a plurality of cameras installed at different heights in a vehicle. step. Identifying, by a processor, an object included in a 2D image captured by a camera having the highest priority based on a priority among the plurality of cameras, and capturing by a processor the camera having the highest priority It may be a computer program recorded on a recording medium to execute the step of transmitting the 2D image and information about the identified object to the learning data generating device.

More specifically, the processor may include an Application-Specific Integrated Circuit (ASIC), other chipsets, logic circuits, and/or data processing devices. The memory may include read-only memory (ROM), random access memory (RAM), flash memory, memory cards, storage media, and/or other storage devices. The transceiver 160 may include a baseband circuit for processing wired/wireless signals. The input/output device 165 includes an input device such as a keyboard, a mouse, and/or a joystick, and a Liquid Crystal Display (LCD), an Organic LED (OLED), and/or a liquid crystal display (LCD). Alternatively, an image output device such as an active matrix OLED (AMOLED) may include a printing device such as a printer or a plotter.

When the embodiments included in this specification are implemented as software, the above-described method may be implemented as a module (process, function, etc.) that performs the above-described functions. A module resides in memory and can be executed by a processor. The memory may be internal or external to the processor and may be coupled with the processor in a variety of well-known means.

Each component may be implemented by various means, eg, hardware, firmware, software, or a combination thereof. In the case of hardware implementation, one embodiment of the present invention includes one or more Application Specific Integrated Circuits (ASICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DSPDs), Programmable Logic Devices (PLDs), FPGAs ( Field Programmable Gate Arrays), processors, controllers, microcontrollers, microprocessors, etc.

In addition, in the case of implementation by firmware or software, an embodiment of the present invention is implemented in the form of a module, procedure, function, etc. that performs the functions or operations described above, and is stored on a recording medium readable through various computer means. can be recorded. Here, the recording medium may include program commands, data files, data structures, etc. alone or in combination. Program instructions recorded on the recording medium may be those specially designed and configured for the present invention, or those known and usable to those skilled in computer software. For example, recording media include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CD-ROMs (Compact Disk Read Only Memory) and DVDs (Digital Video Disks), floptical It includes hardware devices specially configured to store and execute program instructions, such as magneto-optical media, such as a floptical disk, and ROM, RAM, flash memory, and the like. Examples of program instructions may include high-level language codes that can be executed by a computer using an interpreter or the like as well as machine language codes generated by a compiler. These hardware devices may be configured to operate as one or more pieces of software to perform the operations of the present invention, and vice versa.

With the following configuration, the de-identification processing unit may de-identify a part of the 2D image in correspondence to the type of object identified by the object identification unit.

Specifically, the de-identification processing unit may perform de-identification processing on a region corresponding to the de-identification processing region previously assigned to the 3D model when the object identification unit sets the object region using the 3D model.

To this end, the de-identification processing unit may extract an edge of the 2D image, and search a database for a de-identification rule corresponding to the type of object identified by the object identification unit and the pattern of the edge extracted from the 2D image. When a de-identification processing rule is retrieved from the database, the de-identification processing unit may de-identify a part of the 2D image according to the retrieved de-identification processing rule. In this case, the de-identification processing rule is distributed by the learning data generating device, and is a rule enumerating edge patterns for de-identification areas existing in the object so as to determine areas to be de-identified within the 2D image.

The de-identification processing unit requests an update to the de-identification processing rule by including the 2D image, the type of object identified from the 3D point cloud data, and the edge extracted from the 2D image when the de-identification processing rule is not retrieved from the database. messages can be created. When the update request message is generated, the de-identification processing unit may transmit the generated update request message to the learning data generating device through the communication unit.

On the other hand, in performing de-identification processing, the de-identification processing unit performs de-identification processing, the boundary line of the area to be de-identified and the boundary line of the enclosure by the edge extracted from the 2D image is a preset minimum separation distance When adjacent to the inside, the size of the region to be de-identified may be reduced so that the boundary of the region to be de-identified is separated from the boundary of the closed region by a minimum distance or more. In this case, the minimum separation distance may be the minimum number of pixels positioned between the two boundary lines.

In this way, the de-identification processing unit minimizes the effect of de-identification processing of personal information on the annotation work by ensuring that the boundary of the area to be de-identified and the edge of the 2D image do not overlap each other and are separated by a minimum distance or more. can

Meanwhile, the de-identification processing unit as described above may perform de-identification processing by blurring a part of the 2D image. In this case, since the visibility of the 2D image that is subject to de-identification processing may change according to the moving speed of the vehicle, the de-identification processing unit sets the intensity of blurring for de-identification processing to the moving speed of the vehicle in which the camera is installed. It can also be set in inverse proportion to .

Also, the de-identification processing unit may crop an area to be subjected to de-identification processing before performing blurring processing, and then include the cropped image to generate restored data.

Hereinafter, a method for refining raw data for artificial intelligence learning according to importance evaluation according to various embodiments of the present invention as described above and a computer program recorded on a recording medium for executing the method will be described in detail.

In the present invention, prior to the data collection server performing de-identification processing, etc., raw data of low importance can be refined. To this end, the data collection server may evaluate the importance of each raw data, and the importance of the raw data may basically include whether an object to be an artificial intelligence learning target is included or not, and photographing environment information.

In addition, the data collection server may calculate the similarity between an object included in the raw data and an object to be an artificial intelligence learning target, and remove raw data having a similarity lower than a threshold to perform purification.

In addition, the data collection server may perform purification by removing raw data that is different from the environment related to artificial intelligence learning based on information (GPS, weather information, etc.) on the shooting time of raw data.

As described above, although preferred embodiments of the present invention have been disclosed in the present specification and drawings, it is in the technical field to which the present invention belongs that other modified examples based on the technical spirit of the present invention can be implemented in addition to the embodiments disclosed herein. It is self-evident to those skilled in the art. In addition, although specific terms have been used in the present specification and drawings, they are only used in a general sense to easily explain the technical content of the present invention and help understanding of the present invention, but are not intended to limit the scope of the present invention. Accordingly, the foregoing detailed description should not be construed as limiting in all respects and should be considered illustrative. The scope of the present invention should be selected by reasonable interpretation of the appended claims, and all changes within the equivalent scope of the present invention are included in the scope of the present invention.

Claims (5)

As a data purification method of a data collection device for machine learning of artificial intelligence (AI),
Collecting, by a learning data collection device, data using a sensor;
refining the collected data by the learning data collection device; and
Evaluating, by the learning data collection device, importance of the collected data; and
Transmitting, by the learning data collection device, refined data to a data collection server;
The purification step is
Raw data purification method for artificial intelligence learning according to importance evaluation, characterized in that the similarity between the object included in the data and the object to be artificial intelligence learning is calculated and performed.
According to claim 1,
The purification step is
A method for refining raw data for artificial intelligence learning according to importance evaluation, characterized in that raw data having a similarity between an object included in the data and an object to be artificial intelligence learning is less than a threshold is removed and refined.
According to claim 2,
The purification step is
For artificial intelligence learning according to importance evaluation, characterized in that the learning data collection device performs purification by removing raw data that is different from the environment related to artificial intelligence learning based on the information about the time of shooting of the raw data. Raw data cleaning method.
According to claim 3,
The learning data collection device,
Based on the meta data related to the collection time of the raw data, the raw data purification rate is controlled, and based on the collected meta data, the degree of influence of the raw data on artificial intelligence learning is calculated, characterized in that the importance level Raw data purification method for artificial intelligence learning according to evaluation.
memory;
transceiver; and
In combination with a computing device configured to include a processor for processing instructions resident in the memory,
Collecting, by the processor, data using a sensor;
Step, by the processor, refining the collected data; and
Evaluating, by the processor, importance of the collected data; and
Executing, by the processor, transmitting the refined data to a data collection server;
The purification step is
A computer program recorded on a recording medium, characterized in that the similarity between the object included in the data and the object to be artificial intelligence learning is calculated and performed.