Residual W-shape network (ResWnet) for dual-energy cone-beam CT imaging

Abstract

Deep learning has achieved great success in many medical imaging tasks without explicit solutions. In this work, learning method was applied to dual-energy cone-beam CT imaging. We proposed a Residual W-shape Network (ResWnet). ResWnet consists of three modules: scatter correction module 𝒮, material decomposition module ℳ, decomposition denoising module 𝒟. Both 𝒮 and 𝒟 use ResUnet architecture, and this lightweight model fuses multi-level features, achieving satisfied performance with a small number of parameters. 𝒮 acts on dual-energy attenuation projections to reduce the scatter contaminations, and 𝒟 acts on material composition projections to suppress the noise. ℳ links the modules 𝒮 and 𝒟, and is used for domain transform from attenuation projections to material projections. This process could be approximated by polynomials with pre-calibrated parameters, that is, ℳ is a known operator in proposed network with no trainable parameters. This helps to reduce model parameters and improve the performance with small training dataset. Using public head CT dataset, we simulated dual-energy cone-beam CT projections and material projections. Proposed ResWnet was trained, validated and tested on this simulated dataset, verifying its effectiveness in projection-domain scatter correction and low-noise decomposition.

Using ResWnet for semantic segmentation of active wildfires from Landsat-8 imagery

“…However, this shouldn't be confused with a residual W-Net. 23 W-Nets, although W-shaped, have a very different architecture compared with the ResWnet. With W (Double U)-Nets, the first "U" serves as the encoder, and the second "U" serves as the decoder, and each "U" has 4 down-and 4 up-convolutions.…”

Feasibility of dual‐energy CBCT material decomposition in the human torso with 2D anti‐scatter grids and grid‐based scatter sampling

“…Robust solutions for scatter mitigation are needed to enable DE CBCT imaging of the human torso. Several methods have been investigated, including radiographic anti-scatter grids, 16,24 scatter correction methods, 11 deep learning methods, 25,26 and primary modulation methods. [27][28][29] This study proposes and investigates a new approach for suppressing scatter in DE CBCT imaging of the human torso.…”