Demo abstract: DNN-based SLAM Tracking Error Online Estimation https://sites.duke.edu/tianyihu/files/2023/07/DemoVideoWithCaption.mp4 Figure 1: Our demo system architecture, which showcases DeepSEE evaluated in a virtual environment to calculate the estimated pose and corresponding pose error at run time. Figure 2: DeepSEE in action. Demo participants interact with a virtual environment using a keyboard; DeepSEE estimates the pose error of visual SLAM running in the virtual environment.