Learning from Observations Using a Single Video Demonstration and Human Feedback