Adaptive Refinement of Segmented Object Contour Based on the Brightness of Neighboring Pixels Using the Ensemble Method

image_print
DOI https://doi.org/10.15407/pmach2024.04.073
Journal Journal of Mechanical Engineering – Problemy Mashynobuduvannia
Publisher Anatolii Pidhornyi Institute of Power Machines and Systems
of National Academy of Science of Ukraine
ISSN  2709-2984 (Print), 2709-2992 (Online)
Issue Vol. 27, no. 4, 2024 (December)
Pages 73-78
Cited by J. of Mech. Eng., 2024, vol. 27, no. 4, pp. 73-78

 

Author

Vladyslav D. Koniukhov, Anatolii Pidhornyi Institute of Power Machines and Systems of NAS of Ukraine (2/10, Komunalnykiv str., Kharkiv, 61046, Ukraine), e-mail: riggelllll@gmail.com, ORCID: 0009-0007-0256-1388

 

Abstract

Improving the accuracy of computer vision algorithms plays a significant role in the tasks of medical image segmentation. After all, determining the boundaries of objects is a difficult task when using medical images, and especially X-ray images. The use of X-ray images in segmentation tasks is a complex process, since these images themselves can have a sufficient amount of noise and artifacts. Classical segmentation methods face significant challenges when segmenting X-ray images where there are objects with fuzzy boundaries. To solve such tasks, it is suggested to use segmentation with the help of machine learning, and to increase the accuracy of determining the boundaries of objects, it is necessary to use adaptive approaches. This paper proposes a new method to improve the accuracy of X-ray image segmentation, which analyzes the neighboring pixels of each contour element and adaptively reshapes it if necessary, and then combines all predictions using an ensemble method, which improves the previous version of the contour. The method was able to demonstrate an improvement in the quality of image segmentation on three datasets with different complexity of structures. Improvements in object boundary accuracy were obtained for all three sets.

 

Keywords: machine learning, neural networks, deep learning, image segmentation, medical image analysis.

 

Full text: Download in PDF

 

References

  1. Minaee, S., Boykov, Y., Porikli, F., Plaza, A., Kehtarnavaz, N., & Terzopoulos, D. (2022). Image segmentation using deep learning: A survey. IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 44, no. 7, pp. 3523–3542. https://doi.org/10.1109/TPAMI.2021.3059968.
  2. Zhu, C., Zhang, X., Li, Y., Qiu, L., Han, K., & Han, X. (2022). SharpContour: A contour-based boundary refinement approach for efficient and accurate instance segmentation. 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (New Orleans, LA, USA), pp. 4382–4391. https://doi.org/10.1109/CVPR52688.2022.00435.
  3. Aouat, S., Ait-hammi, I., & Hamouchene, I. (2021). A new approach for texture segmentation based on the Gray Level Co-occurrence Matrix. Multimedia Tools and Applications, vol. 80, pp. 24027–24052. https://doi.org/10.1007/s11042-021-10634-4.
  4. Ma, D., Liao, Q., Chen, Z., Liao, R., & Ma, H. (2019). Adaptive local-fitting-based active contour model for medical image segmentation. Signal Processing: Image Communication, vol. 76, pp. 201–213. https://doi.org/10.1016/j.image.2019.05.006.
  5. Judah, A., Hu, B., & Wang, J. (2014). An algorithm for boundary adjustment toward multi-scale adaptive segmentation of remotely sensed imagery. Remote Sensing, vol. 6, iss. 5, pp. 3583–3610. https://doi.org/10.3390/rs6053583.
  6. Bransby, K. M., Bajaj, R., Ramasamy, A., Çap, M., Yap, N., Slabaugh, G., Bourantas, C., & Zhang, Q. (2024). POLYCORE: Polygon-based contour refinement for improved intravascular ultrasound segmentation. Computers in Biology and Medicine, vol. 182, article 109162. https://doi.org/10.1016/j.compbiomed.2024.109162.
  7. Zia, H., Soomro, S., & Choi, K. N. (2024). Image segmentation using bias correction active contours. IEEE Access, vol. 12, pp. 60641–60655. https://doi.org/10.1109/ACCESS.2024.3391052.
  8. Mohammed, A. & Kora, R. (2023). A comprehensive review on ensemble deep learning: Opportunities and challenges. Journal of King Saud University – Computer and Information Sciences, vol. 35, iss. 2, pp. 757–774. https://doi.org/10.1016/j.jksuci.2023.01.014.
  9. VinDr-SpineXR: An open dataset for spinal lesions detection and classification from radiographs. Vinbigdata: official site. Retrieved from https://vindr.ai/datasets/spinexr.
  10. Jaeger, S., Karargyris, A., Candemir, S., Folio, L., Siegelman, J., Callaghan, F., Xue, Z., Palaniappan, K., Singh, R. K., Antani, S., Thoma, G., Wang, Y. X., Lu, P. X., & McDonald, C. J. (2014). Automatic tuberculosis screening using chest radiographs. IEEE Transactions on Medical Imaging, vol. 33, iss. 2, pp. 233–245. https://doi.org/10.1109/TMI.2013.2284099.
  11. Candemir, S., Jaeger, S., Palaniappan, K., Musco, J. P., Singh, R. K., Xue, Z., Karargyris, A., Antani, S., Thoma, G., & McDonald, C. J. (2014). Lung segmentation in chest radiographs using anatomical atlases with nonrigid registration. IEEE Transactions on Medical Imaging, vol. 33, iss. 2, pp. 577–590. https://doi.org/10.1109/TMI.2013.2290491.

 

Received 26 September 2024

Published 30 December 2024