The 3D EdgeRunner Pipeline: a novel shape-based analysis for neoplasms characterization - Archive ouverte HAL Access content directly
Conference Papers Year :

The 3D EdgeRunner Pipeline: a novel shape-based analysis for neoplasms characterization

Fernando Yepes-C
  • Function : Author
Rebecca Johnson
Yi Lao
  • Function : Author
  • PersonId : 921707
Darryl Hwang
  • Function : Author
Felix Yap
  • Function : Author
Desai Bushan
  • Function : Author
Phillip Cheng
  • Function : Author
Inderbir Gill
  • Function : Author
Vinay Duddalwar
  • Function : Author

Abstract

The characterization of tumors after being imaged is currently a qualitative process performed by skilled professionals. If we can aid their diagnosis by identifying quantifiable features associated with tumor classification, we may avoid invasive procedures such as biopsies and enhance efficiency. The aim of this paper is to describe the 3D EdgeRunner Pipeline which characterizes the shape of a tumor. Shape analysis is relevant as malignant tumors tend to be more lobular and benign ones tare generally more symmetrical. The method described considers the distance from each point on the edge of the tumor to the centre of a synthetically created field of view. The method then determines coordinates where the measured distances are rapidly changing (peaks) using a second derivative found by five point differentiation. The list of coordinates considered to be peaks can then be used as statistical data to compare tumors quantitatively. We have found this process effectively captures the peaks on a selection of kidney tumors.
Fichier principal
Vignette du fichier
ERP-DiffMethod_Spie2015.pdf (920.21 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-01710708 , version 1 (15-05-2018)

Identifiers

Cite

Fernando Yepes-C, Rebecca Johnson, Yi Lao, Darryl Hwang, Julie Coloigner, et al.. The 3D EdgeRunner Pipeline: a novel shape-based analysis for neoplasms characterization. SPIE Medical Imaging 2016, Feb 2016, San Diego, CA, United States. ⟨10.1117/12.2217238⟩. ⟨hal-01710708⟩
46 View
71 Download

Altmetric

Share

Gmail Facebook Twitter LinkedIn More