Abstract

Gait recognition has emerged as a vital biometric technique for unobtrusive human identification in surveillance, healthcare, and behavioral analytics. However, achieving high accuracy under real-world variations such as walking speed, clothing, and viewpoint changes remains a significant challenge. This study proposes an advanced gait recognition framework that combines a Convolutional Neural Network (CNN) for spatial feature extraction with a Long Short-Term Memory (LSTM) network for modeling temporal dynamics. To address the limitations of manual hyperparameter tuning, the Hippopotamus Optimization Algorithm (HOA); a bio-inspired metaheuristic is integrated to optimize key parameters such as learning rate, filter size, LSTM units, and dropout rate. The model is evaluated on the TUM GAID dataset, enco­­­­­mpassing diverse gait variations. Experimental results demonstrate that the HOA-optimized CNN-LSTM architecture significantly outperforms baseline and state-of-the-art methods in terms of recognition accuracy, Genuine Acceptance Rate (GAR), and Equal Error Rate (EER). The proposed framework exhibits superior robustness and convergence speed, affirming the efficacy of metaheuristic-driven optimization in deep learning-based gait recognition systems.

Keywords

  • Gait Biometrics
  • Convolutional Neural Networks (CNN)
  • Long Short-Term Memory (LSTM)
  • Metaheuristic Optimization
  • Hippopotamus Optimization Algorithm

References

  1. 1. A. Ahmed, M. Tariq, and R. Khan, “Biometric Recognition Systems: A Review on Modalities, Challenges, and Applications,” IEEE Access, vol. 10, pp. 12837–12856, 2022.
  2. 2. X. Zhao, Y. Yang, and H. Liu, “Advances in Gait Recognition: A Survey,” ACM Comput. Surv., vol. 54, no. 5, pp. 1–36, Sep. 2022.
  3. 3. J. Chen, H. Wang, and Z. Wu, “View-Invariant Gait Recognition: A Survey,” IEEE Trans. Biometrics, Behavior, and Identity Science, vol. 5, no. 1, pp. 1–19, Jan. 2023.
  4. 4. M. Rajput, S. Agrawal, and A. Tiwari, “Model-Based Gait Recognition Approaches: Challenges and Trends,” in Proc. Int. Conf. Adv. Comput., 2022, pp. 231–236.
  5. 5. M. Alsolai and M. Albahli, “Deep Learning for Gait Analysis Using CNN-LSTM Models,” IEEE Access, vol. 12, pp. 482–490, 2024.
  6. 6. L. Lu et al., “Spatiotemporal Gait Recognition Using ResNet and LSTM Fusion,” Pattern Recognition Letters, vol. 165, pp. 134–142, 2023.
  7. 7. S. Wang et al., “Transfer Learning in Gait Recognition: Cross-View and Cross-Condition Solutions,” Neurocomputing, vol. 520, pp. 133–145, 2023.
  8. 8. Z. Chen, Y. Li, and W. Xu, “Hyperparameter Optimization for Gait Recognition Using Metaheuristic Algorithms,” Applied Soft Computing, vol. 140, p. 110743, 2024.
  9. 9. J. Ahmed, N. Sharma, and A. Rauf, “LSTM Networks for Biometric Gait Analysis: A Review,” IEEE Reviews in Biomedical Engineering, vol. 16, pp. 25–38, 2023.
  10. 10. S. Liang, H. Zhou, and L. Yang, “Hybrid CNN-LSTM for Gait Classification Using Wearable Sensors,” Sensors, vol. 25, no. 2, pp. 2293–2310, 2025.
  11. 11. R. Kumar, P. Goyal, and D. Singh, “Bi-LSTM-Based Gait Recognition in Smart Surveillance,” Computers, Materials & Continua, vol. 77, no. 3, pp. 3981–3997, 2023.
  12. 12. T. Chen and M. Zhang, “CNN-LSTM Hybrid Architecture for Gait Event Detection,” Computational Intelligence and Neuroscience, vol. 2023, Article ID 2732546, 2023.
  13. 13. H. Wu, C. Li, and J. Zhang, “Challenges in Optimizing Deep Learning Models for Gait Recognition,” Expert Systems with Applications, vol. 226, p. 119648, 2023.
  14. 14. A. Wang, Y. Zeng, and Q. Zhao, “Metaheuristic Algorithms for Neural Network Optimization: A Survey,” IEEE Trans. Neural Networks Learn. Syst., vol. 35, no. 2, pp. 1–18, 2023.
  15. 15. H. Ahmed, R. Iqbal, and S. Khan, “Hippopotamus Optimization Algorithm for Image Classification,” Expert Systems with Applications, vol. 209, p. 118311, 2023.
  16. 16. Y. Liang, B. Lin, and M. Liu, “HOA: A Novel Metaheuristic for Biomedical Image Processing,” Applied Intelligence, vol. 65, no. 4, pp. 2517–2530, 2025.
  17. 17. TUM GAID Dataset: Gait from Audio, Image and Depth. Accessed: Jan. 2025. [Online]. Available: https://vision.in.tum.de/data/datasets/gait
  18. 18. X. Zhao, Y. Yang, and H. Liu, “Advances in Gait Recognition: A Survey,” ACM Comput. Surv., vol. 54, no. 5, pp. 1–36, Sep. 2022.
  19. 19. A. Ahmed, M. Tariq, and R. Khan, “Biometric Recognition Systems: A Review,” IEEE Access, vol. 10, pp. 12837–12856, 2022.
  20. 20. J. Chen, H. Wang, and Z. Wu, “View-Invariant Gait Recognition: A Survey,” IEEE Trans. Biometrics, Behavior, and Identity Science, vol. 5, no. 1, pp. 1–19, Jan. 2023.
  21. 21. S. N. Singh, “Gait as a Biometric: Recognition from Temporal Motion Patterns,” Journal of Information Security, vol. 9, pp. 1–14, 2021.
  22. 22. M. Rajput, S. Agrawal, and A. Tiwari, “Model-Based Gait Recognition,” in Proc. Int. Conf. Adv. Comput., 2022, pp. 231–236.
  23. 23. K. He, X. Zhang, S. Ren, and J. Sun, “Deep Residual Learning for Image Recognition,” in Proc. CVPR, 2016, pp. 770–778.
  24. 24. J. Ahmed, N. Sharma, and A. Rauf, “LSTM Networks for Biometric Gait Analysis,” IEEE Reviews in Biomedical Engineering, vol. 16, pp. 25–38, 2023.
  25. 25. M. Alsolai and M. Albahli, “Deep Learning for Gait Analysis Using CNN-LSTM Models,” IEEE Access, vol. 12, pp. 482–490, 2024.
  26. 26. R. Kumar, P. Goyal, and D. Singh, “Bi-LSTM-Based Gait Recognition in Smart Surveillance,” Computers, Materials & Continua, vol. 77, no. 3, pp. 3981–3997, 2023.
  27. 27. Z. Chen, Y. Li, and W. Xu, “Hybrid CNN-LSTM Models for Gait Recognition,” Applied Soft Computing, vol. 140, p. 110743, 2024.
  28. 28. H. Wu, C. Li, and J. Zhang, “Hyperparameter Optimization in Deep Learning,” Expert Systems with Applications, vol. 226, p. 119648, 2023.
  29. 29. D. P. Kingma and J. Ba, “Adam: A Method for Stochastic Optimization,” in Proc. ICLR, 2015.
  30. 30. J. H. Holland, Adaptation in Natural and Artificial Systems, University of Michigan Press, 1975.
  31. 31. J. Kennedy and R. Eberhart, “Particle Swarm Optimization,” in Proc. IEEE Int. Conf. Neural Networks, 1995, pp. 1942–1948.
  32. 32. S. Mirjalili and A. Lewis, “The Whale Optimization Algorithm,” Adv. Eng. Softw., vol. 95, pp. 51–67, 2016.
  33. 33. X. S. Yang, “Firefly Algorithms for Multimodal Optimization,” in Stochastic Algorithms: Foundations and Applications, 2009, pp. 169–178.
  34. 34. H. Ahmed, R. Iqbal, and S. Khan, “Hippopotamus Optimization Algorithm for Neural Network Training,” Expert Systems with Applications, vol. 209, p. 118311, 2023.
  35. 35. Y. Liang, B. Lin, and M. Liu, “HOA for Biomedical Image Classification,” Applied Intelligence, vol. 65, no. 4, pp. 2517–2530, 2025.
  36. 36. Y. LeCun, Y. Bengio, and G. Hinton, "Deep learning," Nature, vol. 521, no. 7553, pp. 436–444, 2015.
  37. 37. M. F. Abed, S. Mirjalili, and A. H. Gandomi, "Hippopotamus Optimization Algorithm: A novel nature-inspired optimizer," Expert Systems with Applications, vol. 167, p. 114133, Mar. 2021.
  38. 38. S. Mirjalili, "Evolutionary algorithms and neural networks," Studies in Computational Intelligence, vol. 780, pp. 1–15, 2019.
  39. 39. M. H. Amiri, N. M. Hashjin, M. Montazeri, S. Mirjalili, and N. Khodadadi, "Hippopotamus optimization algorithm: A novel nature-inspired optimization algorithm," Sci. Rep., vol. 14, p. 5032, 2023. [Online]. Available: research.torrens.edu.au.
  40. 40. J. Chen, Y. Zhang, and Q. Li, "Gait recognition model using CNNs for feature extraction and shallow classifiers," Journal of Biometric Studies, vol. 45, no. 3, pp. 245–257, 2022.
  41. 41. W. Lin et al., “AI Models for Cross-View Gait Recognition,” IEEE Transactions on Image Processing, vol. 32, pp. 2194–2207, 2023.
  42. 42. S. Wang et al., “Transfer Learning in Gait Recognition,” Neurocomputing, vol. 520, pp. 133–145, 2023.
  43. 43. M. Gupta et al., “Sensor-Based Gait Analysis Using Wearable Devices,” IEEE Sensors Journal, vol. 21, no. 15, pp. 17539–17548, 2021.
  44. 44. TUM-GAID Dataset. [Online]. Available: https://vision.in.tum.de/data/datasets/gait
  45. 45. D. Zhu, X. Huang, X. Wang, B. Yang, B. He, W. Liu, and B. Feng, “Multi-scale context-aware network with transformer for gait recognition,” arXiv, Apr. 2022. Available: https://arxiv.org/abs/2204.03270?utm
  46. 46. J. Yue, Z. Mei, K. Ivanov, Y. Li, T. He, and H. Zeng, “Gait recognition by sensing insole using a hybrid CNN-attention-LSTM network,” in Proc. Int. Conf. Biometric Recognition (CCBR), Nov. 2022, pp. 484–492. Available: https://link.springer.com/chapter/10.1007/978-3-031-20233-9_49?utm
  47. 47. L.-F. Shi, Z.-Y. Liu, K.-J. Zhou, Y. Shi, and X. Jing, “Novel deep learning network for gait recognition using multimodal inertial sensors,” Sensors, vol. 23, no. 2, p. 849, Jan. 2023. Available: https://www.mdpi.com/1424-8220/23/2/849?utm
  48. 48. “Real-time walking gait terrain classification from foot-mounted inertial measurement unit using convolutional LSTM neural network,” Expert Systems with Applications, vol. 203, p. 117306, Oct. 2022. Available: https://www.sciencedirect.com/science/article/pii/S0957417422006698?utm
  49. 49. “Accurate gait recognition with inertial sensors using a new FCN-BiLSTM architecture,” Computers and Electrical Engineering, vol. 104, p. 108428, Dec. 2022. Available: https://www.sciencedirect.com/science/article/abs/pii/S0045790622006450?utm
  50. 50. Y. Yang, L. Yun, R. Li, F. Cheng, and K. Wang, “Multi-view gait recognition based on a Siamese vision transformer,” Appl. Sci., vol. 13, no. 4, p. 2273, Feb. 2023. Available: https://www.mdpi.com/2076-3417/13/4/2273?utm
  51. 51. P. Delgado-Santos, R. Tolosana, R. Guest, F. Deravi, and R. Vera-Rodriguez, “Exploring transformers for behavioural biometrics: A case study in gait recognition,” arXiv, Jun. 2022. Available: https://arxiv.org/abs/2206.01441?utm
  52. 52. L. Arshad, A. Jamsrandorj, J. Kim, and K.-R. Mun, “Gait events prediction using hybrid CNN-RNN-based deep learning models through a single waist-worn wearable sensor,” arXiv, Feb. 2022. Available: https://arxiv.org/abs/2203.00503?utm
  53. 53. S. “A multimodal CNN–Transformer network for gait pattern recognition with wearable sensors in weak GNSS scenarios,” Electronics, vol. 14, no. 8, p. 1537, Apr. 2023. Available: https://www.mdpi.com/2079-9292/14/8/1537?utm
  54. 54. “Multimodal gait abnormality recognition using a convolutional neural network–bidirectional LSTM network based on multi-sensor data fusion,” Sensors, vol. 23, no. 22, p. 9101, Nov. 2023.
  55. 55. D. Zhu et al., “Multi-scale context-aware network with transformer for gait recognition,” arXiv preprint arXiv:2204.03270, Apr. 2022.
  56. 56. L.-F. Shi et al., “Novel deep learning network for gait recognition using multimodal inertial sensors,” Sensors, vol. 23, no. 2, p. 849, Jan. 2023.
  57. 57. Y. Liu et al., “DPF-LSTM-CNN hybrid network for gait phase recognition,” Sensors, vol. 23, no. 13, p. 5905, Jul. 2023.
  58. 58. Y. Yang et al., “Multi-view gait recognition based on a Siamese vision transformer,” Applied Sciences, vol. 13, no. 4, p. 2273, Feb. 2023.
  59. 59. P. Delgado-Santos et al., “Exploring transformers for behavioural biometrics: A case study in gait recognition,” arXiv preprint arXiv:2206.01441, Jun. 2022.
  60. 60. A. Kumar et al., “Optimized CNN-LSTM framework for prosthetic gait classification,” in Proc. Int. Conf. Cyber-Physical Systems and Advanced Computing (CPSAC), Springer, 2025.
  61. 61. E. Guillen Pinto, L. Ramirez Lopez, and C. Ramos Linares, "GaHu-Video: Parametrization system for human gait recognition," Mendeley Data, 30-May-2020. [Online]. Available: https://data.mendeley.com/datasets/gprg4s73v4/1. [Accessed: 14-Nov-2024]. DOI: 10.17632/gprg4s73v4.1.