Visual Detection of Zebrafish Larva

Visual detection of key features of zebrafish larva has created a climate of preoccupation among micromanipulation practitioners. As the visual detection is the feedback component from the perspective of control theory, an accurate detection of the region of interest plays a pivotal role in the stability and performance of the whole system. That zebrafish larvae differ from each other significantly poses great challenge to the image processing programs, demanding highly robust algorithms that are able to handle the disturbance not only come from the individual difference of the larvae, but also from the unpredictable spontaneous movement of the experimental object. Meanwhile, in order to achieve high-throughput injection, the visual feedback component must be efficient, meaning that our methods should also be optimized so that they can function in real time. Against this backdrop, a set of image processing techniques are proposed.