University of California, Merced

I devoted myself to launch several projects in the field of computer vision. Specifically, I proposed a generative model built for transferring a story or noise vector into a clip of video; an innovative multi-scale encoder-decoder architecture, combined with short connections and dense information to grasp global information better, to detect salient objects; a creative neural network to solve the low accuracy problem by removing the gap between weak high-resolution features and improper anchor settings for aerial object detection.