DetectorShop: Democratizing Deep Learning Object Detectors
Deep learning object detectors have been demonstrated to be critical components in emerging artificial intelligent applications such as autonomous vehicles, cognitive assistants, and search and rescue drones. Due to the complexity of building deep learning object detectors, a toolchain for lowering the barrier has significant value. This thesis presents an end-to-end system to lower the barrier of building deep learning object detectors. The major contribution of this thesis is the design and implementation of DetectorShop, a desktop-based system that works as a PhotoShop for building object detectors.