Transmission X-ray microscopy(TXM)and electron microscopy(EM)arevery common and useful tools toinvestigate specimen cells.Especially for Soft X-ray tomography(SXT),it can directly image the whole cell up to 10 μm without sectioning,staining,dehydration or plastic embedment.For cell imaging,TEM mesh grids are usually used to carry cell specimens.In this case,the mesh grids will block the specimens at a high tilt angle,which will result in “missing wedge” situation and artifact in the computed tomography(CT)reconstruction.In this paper,a new algorithm–the recovering “missing wedge” discrete algebraic reconstruction technique(rmwDART)was proposed to improve the “missing wedge” problem.It can acquire a high-quality reconstruction of multi-gray level material without prior knowledge of the material because over-segmentation is used to segment the image.This indicates that the image segmentation process doesnt need the prior knowledge of the materials,neither the material components number nor the materials value,because the over-segmentation strategy doesnt need these prior knowledge to segment precisely and just overly segments the image.By over-segmentation,boundary extraction and mathematical morphological operations,“missing wedge” artifact areas are located.By updating the located regions withthe iterativereconstructionmethod in the iteration process,high-quality resulting reconstructions can be obtained.The projections of E.coli cells were acquired from the soft X-ray Nano-CT experiments at BL07W beamline at the National Synchrotron Radiation Laboratory(NSRL)in Hefei,China.One slice of the reconstruction of E.coli cell was acquired through the rmwDART algorithm.