본문 바로가기
KERD
논문KCI 등재

불균형한 클래스 데이터에서 오버샘플링 및 3차원 보행 분석을 통한 BMI 분류 정확도 개선

Improving BMI Classification Accuracy with Oversampling and 3-D Gait Analysis on Imbalanced Class Data

1

조회

0

다운로드

0

인용

0

참고문헌

키워드

주제분류컴퓨터학

초록

본 연구에서는 3차원 보행 데이터 기반 체질량지수(Body Mass Index, BMI) 추정 기술의 분류 정확도를향상시킬 수 있는 방법을 제안한다. BMI 추정 기술에 관해 기존 연구에서는 BMI 분류 정확도가 약60%에 불과했다. 이에 본 연구에서는 먼저 BMI 분류 정확도가 낮았던 원인을 규명한다. 본 연구의분석 결과에 따르면, 그 원인은 보행 데이터 세트의 클래스 불균형(Class Imbalance) 문제를 해결하기위해 언더샘플링(Undersampling) 기법을 사용한 것에 있었다. 이에 본 연구에서는 언더샘플링 기법 대신오버샘플링(Oversampling) 기법을 적용해 클래스 불균형 문제를 해결하는 것을 제안한다. 또한, 보행데이터 기반 BMI 추정 기술에서 인체 측정학(Anthropometric) 특징과 시공간적(Spatiotemporal) 특징의유용성을 재입증한다. 기존 연구에서는 언더샘플링 기법이 적용된 상태에서 인체 측정학 특징과 시공간적특징의 유용성이 평가됐고, 두 특징을 함께 사용하면 단독으로 사용했을 때보다 BMI 추정 성능이 낮아진다고 보고됐다. 하지만 본 연구 결과에 따르면, 두 특징을 함께 사용하고 오버샘플링 기법을 적용했을때 BMI 추정 문제에서 92.92%의 정확도로 SOTA(State-Of-The-Art) 성능을 달성하는 것을 보인다.

Abstract

In this study, we propose a method to improve the classification accuracy of body mass index (BMI) estimation techniques based on three-dimensional gait data. In previous studies on BMI estimation techniques, the classification accuracy was only about 60%. In this study, we identify the reasons for the low BMI classification accuracy. According to our analysis, the reason is the use of the undersampling technique to address the class imbalance problem in the gait dataset. We propose applying oversampling instead of undersampling to solve the class imbalance issue. We also demonstrate the usefulness of anthropometric and spatiotemporal features in gait data-based BMI estimation techniques. Previous studies evaluated the usefulness of anthropometric and spatiotemporal features in the presence of undersampling techniques and reported that their combined use leads to lower BMI estimation performance than when using either feature alone. However, our results show that using both features together and applying an oversampling technique achieves state-of-the-art performance with 92.92% accuracy in the BMI estimation problem.

참고문헌 (52)

  1. G. Muscogiuri, "Obesity : A gender-view", J. Endocrinol. Invest, 47, pp.299~306 (2024)[DOI]
  2. J. H. Jeong, "DeepHealthNet : Adolescent obesity prediction system based on a deep learning framework", IEEE J. Biomed. Health Inform, 28(4), pp.2282~2293 (2024)[DOI]
  3. Y. Han, "Trends in obesity, leisure-time physical activity, and sedentary behavior in Korean adults : Korea national health and nutritional examinations survey from 2014 to 2021", PLoS One, 19(1), pp.1~15 (2024)[DOI]
  4. D. Mohajan, "Body mass index(bmi)is a popular anthropometric tool to measure obesity among adults", J. Innov. Med. Res, 2(4), pp.25~33 (2023)[DOI]
  5. K. Ricanek, "MORPH : A longitudinal image database of normal adult age-progression", Proc. 7th Int. Conf. Automatic Face and Gesture Recognit, pp.341~345 (2006)
  6. Q. Peng, "A dual-augmentor framework for domain generalization in 3d human pose estimation", Proc. IEEE, 2024, 2240~2249
  7. Z. Jiang, "Back to optimization : diffusion-based zero-shot 3d human pose estimation", Proc. IEEE, 2024, 6142~6152
  8. S. Mehraban, "MotionAGFormer : Enhancing 3d human pose estimation with a transformer-gcnformer network", Proc. IEEE, 2024, 6920~6930
  9. V. O. Andersson, "Gender and body mass index classification using a Microsoft Kinect sensor", Proc. 28th Int. Florida Artificial Intelligence Res. Society Conf, pp.103~106 (2015)
  10. V. Coetzee, "Deciphering faces : Quantifiable visual cues to weight", Perception, 39(1), pp.51~61 (2010)[DOI]
  11. D. D. Pham, "Body mass index and facial cues in sasang typology for young and elderly persons", Evid. -based Complement Altern. Med, 2011(1), pp.1~9 (2011)[DOI]
  12. L. Wen, "A computational approach to body mass index prediction from face images", Image Vis. Comput, 31(5), pp.392~400 (2013)[DOI]
  13. E. Kocabey, "Face-to-bmi : Using computer vision to infer body mass index on social media", Proc. Int. AAAI Conf. Web and Social Media, pp.572~575 (2017)
  14. P. N. Aarotale, "PatchBMI-Net : Lightweight facial patch-based ensemble for bmi prediction", Proc. IEEE Int. Conf. Bioinformatics and Biomed, pp.4022~4028 (2023)
  15. H. Bipembi, "Calculation of body mass index using image processing techniques", Int. J. Artif. Intell. Mechatron, 4(1), pp.1~7 (2015)
  16. J. D. Amador, "Automatic calculation of body mass index using digital image processing", Proc. 5th Workshop on Engineering Appl, pp.309~319 (2018)
  17. A. Pantanowitz, "Estimation of body mass index from photographs using deep convolutional neural networks", Informatics in Medicine Unlocked, 26, pp.1~8 (2021)[DOI]
  18. S. Kim, "Multi-view body image-based prediction of body mass index and various body part sizes", Proc. IEEE, 2023, 6033~6040
  19. Z. Jin, "Attention guided deep features for accurate body mass index estimation", Pattern Recognit. Lett, 154, pp.22~28 (2022)[DOI]
  20. N. U. Niaz, "Class imbalance problems in machine learning : A review of methods and future challenges", Proc. Int. Conf. Comput. Advan, pp.485~490 (2022)
  21. K. Ghosh, "The class imbalance problem in deep learning", Mach. Learn, 113, pp.4845~4901 (2024)[DOI]
  22. S. Goswami, "A literature survey on various aspect of class imbalance problem in data mining", Multimed. Tools Appl, pp.1~26 (2024)
  23. Y. Chen, "A new oversampling method based on triangulation of sample space", IEEE Trans. Syst. Man Cybern. -Syst, 54(2), pp.774~786 (2024)[DOI]
  24. W. Pei, "A survey on unbalanced classification : How can evolutionary computation help?", IEEE Trans. Evol. Comput, 28(2), pp.353~373 (2024)
  25. Z. A. Huang, "Neural network with a preference sampling paradigm for imbalanced data classification", IEEE Trans. Neural Netw. Learn. Syst, 35(7), pp.9252~9266 (2024)[DOI]
  26. P. Abdelnour, "Comparing the drop vertical jump tracking performance of the Azure Kinect to the Kinect V2", Sensors, 24(12), pp.1~12 (2024)[DOI]
  27. B. Kwon, "Implementation of human action recognition system using multiple Kinect sensors", Proc. Pacific-Rim Conf. Multimed. (PCM), pp.334~343 (2015)
  28. B. Kwon, "An enhanced multi-view human action recognition system for virtual training simulator", Proc. Asia-Pacific Signal Inf. Process. Assoc. Annu. Summit Conf. (APSIPA ASC), 12, pp.1~4 (2016)
  29. B. Kwon, "Implementation of a virtual training simulator based on 360° multi-view human action recognition", IEEE Access, 5, pp.12496~12511 (2017)[DOI]
  30. B. Kwon, "Human skeleton data augmentation for person identification over deep neural network", Appl. Sci, 10(14), pp.1~22 (2020)[DOI]
  31. B. Kwon, "Ensemble learning for skeleton-based body mass index classification", Appl. Sci, 10(21), pp.1~23 (2020)[DOI]
  32. B. Kwon, "Joint swing energy for skeleton-based gender classification", IEEE Access, 9, pp.28334~28348 (2021)[DOI]
  33. B. Kwon, "Optimal camera point selection toward the most preferable view of 3-D human pose", IEEE Trans. Syst. Man Cybern. -Syst, 52(1), pp.533~553 (2022)[DOI]
  34. B. Kwon, "Multi-time window feature extraction technique for anger detection in gait data", Journal of the Korea Society of Computer and Information, 28(4), pp.41~51 (2023)[DOI]
  35. B. Kwon, "Gait-based gender classification using a correlation-based feature selection technique", Journal of the Korea Society of Computer and Information, 29(3), pp.55~66 (2024)[DOI]
  36. B. Kwon, "An improved energy saving scheme in IEEE 802. 16e", Journal of Korean Institute of Information Technology, 10(8), pp.43~51 (2012)
  37. B. Kwon, "A target position decision algorithm based on analysis of path departure for an autonomous path keeping system", Wirel. Pers. Commun, 83, pp.1843~1865 (2015)[DOI]
  38. B. Kwon, "A downlink power control algorithm for long-term energy efficiency of small cell network", Wirel. Netw, 21, pp.2223~2236 (2015)[DOI]
  39. B. Kwon, "Virtual MIMO broadcasting transceiver design for multi-hop relay networks", Digit. Signal Prog, 46, pp.97~107 (2015)[DOI]
  40. B. Kwon, "Framework implementation of image-based indoor localization system using parallel distributed computing", The Journal of Korean Institute of Communications and Information Sciences, 41(11), pp.1490~1501 (2016)[DOI]
  41. B. Kwon, "Iterative interference cancellation and channel estimation in evolved multimedia broadcast multicast system using filter-bank multicarrier-quadrature amplitude modulation", IEEE Trans. Broadcast, 62(4), pp.864~875 (2016)[DOI]
  42. B. Kwon, "Novel error detection algorithm for LZSS compressed data", IEEE Access, 5, pp.8940~8947 (2017)[DOI]
  43. B. Kwon, "Scattered reference symbol-based channel estimation and equalization for FBMC-QAM systems", IEEE Trans. Commun, 65(8), pp.3522~3537 (2017)[DOI]
  44. B. Kwon, "Effective interference nulling virtual MIMO broadcasting transceiver for multiple relaying", IEEE Access, 5, pp.20695~20706 (2017)[DOI]
  45. B. Kwon, "Cross-antenna interference cancellation and channel estimation for MISO-FBMC", QAM-Based eMBMS, 2018, pp.24, 3281~3293
  46. B. Kwon, "Error detection algorithm for Lempel-Ziv-77 compressed data", J. Commun. Netw, 21(2), pp.100~112 (2019)[DOI]
  47. B. Kwon, "Accurate blind Lempel-Ziv-77parameter estimation via 1-D to 2-D data conversion over convolutional neural network", IEEE Access, 8, pp.43965~43979 (2020)[DOI]
  48. B. Kwon, "EDA-78 : A novel error detection algorithm for Lempel-Ziv-78 compressed data", Wirel. Pers. Commun, 111, pp.2177~2189 (2020)[DOI]
  49. B. Kwon, "Toward an online continual learning architecture for intrusion detection of video surveillance", IEEE Access, 10, pp.89732~89744 (2022)[DOI]
  50. B. Kwon, "Identification and restoration of LZ77 compressed data using a machine learning approach", Proc. Asia-Pacific Signal Inf. Process. Assoc. Annu. Summit Conf. (APSIPA ASC), pp.1787~1790 (2018)
  51. B. Kwon, "Accurate path loss prediction based on a neural network ensemble method", Sensors, 24(1), pp.1~20 (2024)[DOI]
  52. B. Kwon, "Path loss prediction using an ensemble learning approach", Journal of the Korea Society of Computer and Information, 29(2), pp.1~12 (2024)[DOI]

관련 논문AI추천

이 간행물의 다른 논문 (30건)