From: Energy enhanced tissue texture in spectral computed tomography for lesion classification
Texture learning | |
Learning the tissue-specific texture from corresponding full-dose image. | |
Image reconstruction | |
Initialize \( \mathcal{X} \) by algebraic reconstruction technique [25]; | |
 Set parameters λ, μ,  β. | |
 While stop criterion is not met: | |
 Setp1: \( {\mathcal{X}}^{t+1}={\mathrm{argmin}}_{\mathcal{X}}{\left\Vert \mathcal{AX}-\mathcal{Y}\right\Vert}_2^2+\beta R\left(\mathcal{X}\right)+\mu {\left\Vert {\mathcal{D}}^t-\mathcal{X}-{\mathcal{V}}^t\right\Vert}_2^2 \); | |
 Step2: \( {\mathcal{D}}^{t+1}={\mathrm{argmin}}_{\mathcal{D}}\lambda {\left|\mathcal{D}\right|}_{\ast }+\mu {\left\Vert \mathcal{D}-{\mathcal{X}}^{t+1}-{\mathcal{V}}^t\right\Vert}_2^2 \); | |
 Step3: \( {\mathcal{V}}^{t+1}={\mathcal{V}}^t+{\mathcal{X}}^{t+1}-{\mathcal{D}}^{t+1} \); | |
 End until the stop criterion is satisfied. |