Skip to main content
EURASIP Journal on Image and Video Processing
Figure 5 | EURASIP Journal on Image and Video Processing

Figure 5

From: Handling missing weak classifiers in boosted cascade: application to multiview and occluded face detection

Figure 5

Differences between a cascade and a McCascade when several object detectors must be created. A rough estimate of the execution time is given for each step of the training process. (a) A new cascade must be trained for each new object detector. Each new training can last several hours to several days. (b) A new object detector can be trained using a McCascade. Each training just lasts several minutes.

Back to article page