Fig. 7From: Semantic context driven language descriptions of videos using deep neural networkSingle loss versus hybrid lossBack to article page