A gesture judgment device includes a reference part detection unit that outputs reference part information indicating a reference part region, a movement extraction unit that outputs movement information indicating a movement region, a reference part disappearance judgment unit that generates a reference part disappearance information, a timing judgment unit that judges whether first timing indicated by the reference part disappearance information and second timing of occurrence of a frame in which the movement region overlaps with the reference part region indicated by the reference part information are synchronized with each other or not and outputs a timing judgment result as the result of the judging, and an operation judgment unit that judges contents of a gesture operation performed by an operator based on the timing judgment result and the movement information.