Deteksi Tungkai Bayi Pada Image Sequence Berbasis Vector Depth Estimation

Faisal Lutfi Afriansyah; Niyalatul  Muna

doi:10.33795/jip.v7i3.665

Faisal Lutfi Afriansyah Politeknik Negeri Jember
Niyalatul Muna Politeknik Negeri Jember

DOI: https://doi.org/10.33795/jip.v7i3.665

Keywords: infant legs detection, Fundamental matrix, vector depth estimation, motion vector, uncalibrated image sequence

Abstract

Image processing in the image sequence for pattern recognition can be a solution for detecting limb movements in infants after surgery, but the camera is not calibrated. So we need the right method solution to be able to detect these conditions. This happens to cameras that are generally not calibrated and do not have the feature to calculate the vector depth for 3D reconstruction. Because to detect and find limb movement depth is needed to be able to do 3D reconstruction, because it is not only based on the x and y parameters but also z so that with the additional parameters it makes it easier to analyze the motion of the motion axis and the motion vector. This paper discusses a method for detecting 2D motion into a 3D-based motion vector by sequencing the image sequence image then finding the point of transfer of the motion frame destination from the frame reference frame by obtaining the depth (depth vector) using the fundamental matrix from the generated motion vector. This method is recommended because it can perform 3D reconstruction from input in the form of 2D image sequences by calculating the intrinsic parameters so that 3D reconstruction can be carried out. So that the limb vector movement in infants that was originally 2D can be reconstructed into 3D based and makes it easier to carry out the analysis because of the additional parameters.

Downloads

Download data is not yet available.

References

D. P. Snow, “Gesture and intonation are " sister systems " of infant communication: Evidence from regression patterns of language development,” Lang. Commun., vol. 59, pp. 180–191, 2017.
M. Hernández-González, R. M. Hidalgo-Aguirre, M. A. Guevara, M. Pérez-Hernández, and C. Amezcua-Gutiérrez, “Observing videos of a baby crying or smiling induces similar, but not identical, electroencephalographic responses in biological and adoptive mothers,” Infant Behav. Dev., vol. 42, pp. 1–10, 2016.
H. F. Rasmussen et al., “Mother-child language style matching predicts children’s and mothers’ emotion reactivity,” Behav. Brain Res., pp. 1–11, 2016.

H. Simonnet et al., “Parents’ behavior in response to infant crying: Abusive head trauma education,” Child Abus. Negl., vol. 38, no. 12, pp. 1914–1922, 2014.
T. Sun et al., “Development of Lower Limb Motion Detection Based on LPMS *,” 2016.
L. Cattani et al., “Author’s Accepted Manuscript Monitoring Infants by Automatic Video Processing: A Unified Approach to Motion Analysis,” Comput. Biol. Med., vol. 80, no. August 2016, pp. 158–165, 2016.
E. M. Yuniarno, M. Hariadi, and M. H. Purnomo, “Point Cloud Registration for a Non-Deformable,” vol. 51, no. 3, pp. 506–514, 2013.
A. Fogelton and W. Benesova, “Eye blink detection based on motion vectors analysis,” Comput. Vis. Image Underst., vol. 148, pp. 23–33, 2016.

Y. Zhang, L. Zhang, C. Sun, and G. Zhang, “Fundamental Matrix Estimation Based on Improved Genetic Algorithm,” 2016 8th Int. Conf. Intell. Human-Machine Syst. Cybern., no. 3, pp. 326–329, 2016.
Uğur;Topay, Tola;Engin, and A. Aydın, “Solving Fundamental Matrix for Uncalibrated Scene,” Engineering.
A. Z. Richard Hartley, Multiple View Geometry, vol. 53, no. 9. 2003.
F. Zhou, C. Zhong, and Q. Zheng, “Method for fundamental matrix estimation combined with feature lines,” Neurocomputing, vol. 160, pp. 300–307, 2015.
M. Yang, Y. Liu, and Z. You, “Estimating the fundamental matrix based on least absolute deviation,” Neurocomputing, vol. 74, no. 17, pp. 3638–3645, 2011.
C. Steger, “Estimating the fundamental matrix under pure translation and radial distortion,” ISPRS J. Photogramm. Remote Sens., vol. 74, pp. 202–217, 2012.
B. Yildirim and H. A. Ilgin, “Motion vector outlier removal using dissimilarity measure,” Digit. Signal Process. A Rev. J., vol. 46, pp. 1–9, 2015.