Improving annotation performance via speech recognition and visual features - 42Papers