Figure 7
From: Event detection in consumer videos using GMM supervectors and SVMs
The comparison of minimum NDCs with TRECVID2011 MED runs. ‘All’ is the fusion of eight low-level features, ‘Visual’ is the fusion of five visual features, ‘Audio’ is MFCC features, and ‘Motion’ is the fusion of two motion features. ‘Median,’ ‘Average,’ and ‘Min’ is the median, average, and minimum (best) of MinNDCs over 60 runs submitted in TRECVID2011 MED task, respectively.