cascade classifier distinguish and recognize similar objects [closed]
I have taken many pictures.The pictures describe different pose of a car.Can I use the Real Adaboost cascade classifier to recognize specific pose of the car in these pictures? I choosed three images as the positive data and the rest pictures act as the negative data.But the result is bad.
what do you mean by pose of car? Do want to classify it into side view of car and front view of car? Adaboost cascade can be used to recognize one particular pose of car in picture for which the cascade was trained. Also the problem in your case is:Extracting features from three positive images is very unreliable. You need a huge number of positive images dataset and much more number of negative number dataset. Besides, while training the cascade to detect a paticular pose of car in image, make sure all the positive images have same pose.
Thank you! There is only a big car in all the photos. I just want to classify spacial pose of the car. And some positive and negative images are similar. Does that matter? And do I need extract the car from the background? If increase the number of positive images, it may cost lots of time to extract the car from the background.
The cascades are not Classifiers.! They can be used to detect a particular pose of car. eg: If I train a cascade to detect the cars with side view, it will try detecting car with side-view in the test image. The positive images should ideally have object seperated from the background. Besides,more the number of the positive and negative samples, the better detection rate can be expected from the cascade. Typical training of cascade using LBP features wont take more than few hours. For classification, you could probably use SVM.!
Cascade classifiers ARE classifiers. In fact, a bunch of classifiers working together.
Thanks！ What about the size of positive samples? It should be the same with the test image? I think it should be the same with the car in the test image.
I think you should type in cascade classification in the search box and start reading. There have been over 100 topics on this already with quite exhaustive solutions ...
@Lorena GdL: if haar_cascades are classifiers, why cant they be used to classify faces into various types like male/female or classify it into different type of expressions? This made me jump to conclusion that cascades can simply detect presence of certain object but cant be used for classification. Could you explain ?
@tomnjerry: Cascade classifiers are classifiers because they decide whether a certain ROI belongs to the positive class (object) or the negative class (non-object). You're referring to multiclass classification, which can't be achieved (afaik), but without any doubt they perform 2-class classification.
You could use them for male/female classification by using one class as positives and one class as negatives. However you will need alot of data to distinguish between these very similar classes.
@StevenPuttemans: I'm curious about your last comment, I don't really see how that will work. If you use male samples as positive, and supposing you optimally train the system, it will recognise as negative not only female faces but everything else in the world, won't it?