We present a hierarchical imaging reconstruction algorithm for a 3D phase tomography based on the densely extractedfeatures on a multi-band pyramid of convolutional network. By implementing a layer-wise hierarchical machine learningnetwork and combine different bands of information for the imaging retrieval, a more efficient and adaptive learningstrategy is established to enable an accurate reconstruction with fewer training data and improved accuracy. In addition,the distinction of intensity and spectral bands in the feature training process enables bias correction for reconstructionunder varied conditions. In particular, we demonstrate a robust imaging reconstruction for a sparse-view phasetomography application, where spectrally biased phase diffraction and intensity-biased photon noise are bothsuccessfully corrected for.
展开▼
