2013-05-15 19:29:59 -0600 | answered a question | General Cascade Classifier for any type of features It is possible to train OpenCV with other types of features. There is an implementation for LBP features cascades in OpenCV. Basically you need to extend the class FeatureEvaluator and create your MyFeatureEvaluator. The problem, I think, is that the OpenCV code still handles only features that are obtained by comparing, or performing other kinds of operations in rectangular regions inside the images. |
2013-05-15 19:24:46 -0600 | received badge | ● Supporter (source) |
2013-05-15 19:16:46 -0600 | answered a question | cascade classifier training error Are your neg images in gray level mode? This error may be related to number of channels of the images. |
2013-05-15 19:13:24 -0600 | asked a question | How haar like tilted features are represented in xml cascade? Hello everybody! I am trying to flip horizontally the rectangles of xml cascade. For normal rectangles, without tilt, I simply use rect.x = img_width - rect.x - rect.width. But, aparently, that equation does not apply when the rectangle is tilted. When the rectangle is tilted we have the following marks in xml cascade : <tilted>1</tilted> Someone, please, could help to understand to to flip tilted rectangles in the xml cascade? |