Silhouette-based 3D model for zebrafish high-throughput imaging

Good estimations of volume and surface area are important to biological systems measurement. In this paper we develop a 3D reconstruction from evenly sampled axial views in order to enable the volume and surface area measurement. We develop this system for high throughput applications with the zebrafish model system. The VAST BioImager is specifically developed for this purpose and with this system the axial views can be produced. Silhouettes derived from the axial sequence are shape priors which can be directly used to solve the camera calibration problem that is required for the accurate 3D reconstruction. Nonlinear optimisation algorithms have shown to be suitable for the further development of the reconstruction problem. The method proposed in this paper can be included in a measurement pipeline that is used in all kinds of high throughput applications in the zebrafish field. From the 3D reconstruction features can be derived that will contribute to the phenotyping of zebrafish.