KR-20260067691-A - METHOD FOR SELECTING FRAME DATA AND APPARATUS PERFORMING THE SAME

Abstract

A method for selecting frame data and an apparatus for performing the same are disclosed. A method for selecting frame data performed by an electronic device according to one embodiment may include an operation of extracting each feature data of a plurality of frame data from a data set including a plurality of frame data. The method may include an operation of clustering the plurality of frame data into at least one cluster based on the similarity between each feature data of the plurality of frame data. The method may include an operation of selecting at least one frame data from each of the at least one cluster.

Inventors

• 김진환
• 김남곤

Assignees

• 주식회사 에스더블유엠

Dates

Publication Date

20260513

Application Date

20241106

Claims (20)

1. A method for selecting frame data performed by an electronic device, An operation of extracting each feature data of a plurality of frame data from a data set including a plurality of frame data; An operation of clustering the plurality of frame data into at least one cluster based on the similarity between each feature data of the plurality of frame data; and The operation of selecting at least one frame data from each of the above at least one cluster A method including
2. In paragraph 1, The frame data selected from the above at least one cluster is, A method used for training an AI model to obtain information from input frame data.
3. In paragraph 1, The above plurality of frame data are, Frame data captured by sensors mounted on autonomous vehicles A method including
4. In paragraph 1, The operation of extracting each feature data of the plurality of frame data above is, An operation to determine a feature extraction model to be used to extract each feature data of the plurality of frame data based on the characteristics of the plurality of frame data; and The operation of extracting each feature data of the plurality of frame data using a determined feature extraction model. A method including
5. In paragraph 4, The characteristics of the above plurality of frame data are, Information regarding at least one of the complexity, resolution, and noise level of the plurality of frame data above A method including
6. In paragraph 1, The operation of clustering the above plurality of frame data into at least one cluster is, An operation of obtaining similarity between each feature data of the plurality of frame data using a similarity function; and An operation of clustering the plurality of frame data into at least one cluster using a clustering algorithm based on the similarity between each feature data of the plurality of frame data. A method including
7. In paragraph 6, The above similarity function is, A method selected from at least one candidate similarity function based on the dimensions of each feature data of the plurality of frame data.
8. In paragraph 6, The above clustering algorithm is, A method selected from at least one candidate clustering algorithm based on at least one of the number of the plurality of frame data, the dimension of each feature data of the plurality of frame data, and the type of model used to extract each feature data of the plurality of frame data.
9. In paragraph 1, The operation of selecting at least one frame data from each of the above at least one cluster is, The operation of selecting at least one frame data from each of the at least one cluster based on the representative value of each feature data of the frame data included in each of the at least one cluster. A method including
10. In Paragraph 9, The above representative value is, A method determined based on at least one of the mean, median, and mode of each feature data of the frame data included in each of the at least one cluster.
11. In paragraph 1, The operation of selecting at least one frame data from each of the above at least one cluster is, An operation to determine the number of frame data to be selected from each of the at least one cluster based on at least one of the number of frame data included in each of the at least one cluster and at least one of the set parameters of the clustering algorithm used to cluster the plurality of frame data; and The operation of selecting a determined number of frame data from each of the above at least one cluster A method including
12. In a device for selecting frame data, At least one processor; and At least one memory that stores instructions Includes, When the above instructions are executed individually or collectively by the at least one processor, the device is made to perform a plurality of operations, and The above plurality of operations are, An operation of extracting each feature data of a plurality of frame data from a data set including a plurality of frame data; An operation of clustering the plurality of frame data into at least one cluster based on the similarity between each feature data of the plurality of frame data; and The operation of selecting at least one frame data from each of the above at least one cluster A device including
13. In Paragraph 12, The frame data selected from the above at least one cluster is, A device used for training an AI model to obtain information from input frame data.
14. In Paragraph 12, The above plurality of frame data are, Frame data captured by sensors mounted on autonomous vehicles A device including
15. In Paragraph 12, The operation of extracting each feature data of the plurality of frame data above is, An operation to determine a feature extraction model to be used to extract each feature data of the plurality of frame data based on the characteristics of the plurality of frame data; and The operation of extracting each feature data of the plurality of frame data using a determined feature extraction model. A device including
16. In paragraph 15, The characteristics of the above plurality of frame data are, Information regarding at least one of the complexity, resolution, and noise level of the plurality of frame data above A device including
17. In Paragraph 12, The operation of clustering the above plurality of frame data into at least one cluster is, An operation of obtaining similarity between each feature data of the plurality of frame data using a similarity function; and An operation of clustering the plurality of frame data into at least one cluster using a clustering algorithm based on the similarity between each feature data of the plurality of frame data. A device including
18. In Paragraph 17, The above clustering algorithm is, A device selected from at least one candidate clustering algorithm based on at least one of the number of the plurality of frame data, the dimension of each feature data of the plurality of frame data, and the type of model used to extract each feature data of the plurality of frame data.
19. In Paragraph 12, The operation of selecting at least one frame data from each of the above at least one cluster is, The operation of selecting at least one frame data from each of the at least one cluster based on the representative value of each feature data of the frame data included in each of the at least one cluster. Includes, The above representative value is, A device determined based on at least one of the mean, median, and mode of each feature data of frame data included in each of the at least one cluster.
20. In Paragraph 12, The operation of selecting at least one frame data from each of the above at least one cluster is, An operation to determine the number of frame data to be selected from each of the at least one cluster based on at least one of the number of frame data included in each of the at least one cluster and at least one of the set parameters of the clustering algorithm used to cluster the plurality of frame data; and The operation of selecting a determined number of frame data from each of the above at least one cluster A device including

Description

Method for selecting frame data and apparatus for performing the same The following disclosure relates to a method for selecting frame data and an apparatus for performing the same. There are various applications that use AI (artificial intelligence) models (e.g., deep learning models) to process frame data (e.g., images or LiDAR point cloud frame data). For example, in the field of autonomous driving technology, to recognize objects (e.g., vehicles, pedestrians, or traffic signs), recognize lanes, or estimate distances, a vehicle can use a trained model to extract various information (e.g., features) from frame data. In such applications, using an AI model may require updating the model's parameters (e.g., weights or biases) and/or structure through a training process. The data used to train the model is commonly referred to as training data. Since training data significantly impacts the performance of an AI model, it is becoming increasingly important to train the model using high-quality training data. The information described above may be provided as related art for the purpose of aiding understanding of the present disclosure. None of the foregoing information shall be applied as prior art related to the present disclosure. In relation to the description of the drawings, the same or similar reference numerals may be used for identical or similar components. Figure 1 is a diagram illustrating the learning and inference processes of a model that acquires information from frame data. Figure 2 is a diagram illustrating problems that may occur during the training process of a model using a training dataset containing similar frame data. FIG. 3 is a schematic block diagram illustrating a system for selecting frame data according to one embodiment. FIG. 4 is a flowchart illustrating a method for selecting frame data according to one embodiment. FIGS. 5 and 6 are drawings for explaining a method for extracting feature data according to one embodiment. FIGS. 7 to 9 are drawings for explaining a method of clustering frame data according to one embodiment. FIGS. 10 and FIGS. 11 are drawings for explaining a method of selecting frame data from a cluster according to one embodiment. FIG. 12 is a schematic block diagram of an electronic device according to one embodiment. Specific structural or functional descriptions of the embodiments are disclosed for illustrative purposes only and may be modified and implemented in various forms. Accordingly, actual implementations are not limited to the specific embodiments disclosed, and the scope of this specification includes modifications, equivalents, or substitutions included in the technical concept described by the embodiments. Terms such as "first" or "second" may be used to describe various components, but these terms should be interpreted solely for the purpose of distinguishing one component from another. For example, the first component may be named the second component, and similarly, the second component may be named the first component. When it is stated that a component is "connected" to another component, it should be understood that it may be directly connected to or joined to that other component, or that there may be other components in between. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this document, phrases such as “A or B,” “at least one of A and B,” “at least one of A or B,” “A, B or C,” “at least one of A, B and C,” and “at least one of A, B, or C” may each include any one of the items listed together with the corresponding phrase, or all possible combinations thereof. In this specification, terms such as “comprising” or “having” are intended to designate the existence of the described feature, number, step, action, component, part, or combination thereof, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof. Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as generally understood by those skilled in the art. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an ideal or overly formal sense unless explicitly defined in this specification. As used herein, the term "module" may include a unit implemented in hardware, software, or firmware, and may be used interchangeably with terms such as logic, logic block, component, or circuit. A module may be a component formed integrally, or a minimum unit of said component or a part thereof that performs one or more functions. For example, according to one embodiment, a module may be implemented in the form of an application-specific integrated circuit (ASIC). As used in this document, the term "part" refers to a software or hardw