Evaluation of human observer performance on lesion
detectability in single-slice and multislice dedicated breast
cone beam CT images with breast anatomical background

 

M. Han, H. Jang, and J. Baek, Med Phys, 2018. 10

 


 




























Performance comparison of deep learning based
denoising techniques in low-dose CT images

B. Kim, Minah Han, H. Shim, and J. Baek
 CT Meeting, Oral Presentation 2018.06
Fig 7. Pc by human observers (SS), (a) D-DOG ssCHO, and (b) Gabor ssCHO
Fig 8. Pc by human observers (MS1), (a) D-DOG msCHO, and (b) Gabor msCHO
Fig 9. Pc by human observers (MS2), (a) D-DOG msCHO, and (b) Gabor msCHO
Fig 1. The full ROI input image (left) and ROI 1 result images of different denoising methods and reference (right)
Fig 2. Results of quantitative evaluation for test images. The structural similarity index (SSIM) and difference of task
SNR caluated by non-prewhitening with eye filter observer (NPWE) are compared