No abstract

No abstract

Cell imaging and analysis are fundamental to biomedical research because cells are the basic functional units of life. Among different cell-related analysis, cell counting and detection are widely used. In this paper, we focus on one common step of learning-based cell counting approaches: coding the raw dot labels into more suitable maps for learning. Two criteria of coding raw dot labels are discussed, and a new coding scheme is proposed in this paper. The two criteria measure how easy it is to train the model with a coding scheme, and how robust the recovered raw dot labels are when predicting. The most compelling advantage of the proposed coding scheme is the ability to distinguish neighboring cells in crowded regions. Cell counting and detection experiments are conducted for five coding schemes on four types of cells and two network architectures. The proposed coding scheme improves the counting accuracy versus the widely-used Gaussian and rectangle kernels up to 12%, and also improves the detection accuracy versus the common proximity coding up to 14%.

No abstract

Microscopy is widely used for brain research because of its high resolution and ability to stain for many different biomarkers. Since whole brains are usually sectioned for tissue staining and imaging, reconstruction of 3D brain volumes from these sections is important for visualization and analysis. Recently developed tissue clearing techniques and advanced confocal microscopy enable multilayer sections to be imaged without compromising the resolution. However, noticeable structure inconsistence occurs if surface layers are used to align these sections. In this paper, a structure-based intensity propagation method is designed for the robust representation of multilayer sections. The 3D structures in reconstructed brains are more consistent using the proposed methods. Experiments are conducted on 367 multilayer sections from 20 mouse brains. The average reconstruction quality measured by the structure consistence index increases by 45% with the tissue flattening method, and 29% further with the structure-based intensity propagation.