It Takes Two to Tango: Combining Visual and Textual Information for Detecting Duplicate Video-Based Bug Reports