Can you help me about Image classification using SIFT features?I want to classify images based on SIFT features,
1) given a training set of images, extract SIFT from them
2) compute K-Means over the entire set of SIFTs extracted form the training set.
the "K" parameter (the number of clusters) depends on the number of SIFTs that you have for training, but usually is around 500->8000 (the higher, the better).
Now you have obtained K cluster centers.
3) You can compute the descriptor of an image by assigning each SIFT of the image to one of the K clusters. In this way you obtain a histogram of length K.
4)I have 130 images in training set so my training set 130*K dimensional
5)ı want to classify my test images ı have 1 images so my sample is 1*k dimensional. I wrote this code knnclassify(sample,training set,group). I want to classify to 7 group. so knnclassify(sample(1*10),trainingset(130*10),group(7*1)).
the error is : The length of GROUP must equal the number of rows in TRAINING. what can ı do
which library do you use to make your kmeans ? maybe #group mist be equal to k, to indicate in which group your query belong to...
maybe each object in learning set must have label (name of class, in your case [1 to 7])
hi,
go through the link,
http://www.dcs.gla.ac.uk/~srogers/firstcourseml/matlab/chapter5/knnexample.html
Hi,
Perhaps you should review nearest neighbour classification:
https://class.coursera.org/machlearning-001/lecture/203
and try again.
If you want to classify in 7 groups, it means you have to select k=7
you can try the SVM method for classification works better than kNN. Here is the best Library that I have found to use, and the implementation is with PHOW which has the same feature vector as SIFT . Try it out. http://www.vlfeat.org/applications/caltech-101-code.html
thank you for your reply's. I solved this problem but performance is very low so ı want to svm. but ı have 10 class and svm doesn't support multi class so I wrote code multiclass svm(http://www.mathworks.com/matlabcentral/fileexchange/39352-multi-class-svm) but matlab eror is :
Error using seqminopt>seqminoptImpl (line 198)
No convergence achieved within maximum number of iterations.
Error in seqminopt (line 81)
[alphas offset] = seqminoptImpl(data, targetLabels, ...
Error in svmtrain (line 499)
[alpha, bias] = seqminopt(training, groupIndex, ...
Error in multisvm (line 20)
models(k) = svmtrain(TrainingSet,G1vAll);
Error in den (line 57)
results = multisvm(histtrain,grupp,histogramtst);
.This error only kmeans k=>150. I wrote code for k means k=30 and it doesn't give an error.
Try to use Libsvm toolbox. It has Matlab version.
The parameters in SVM can affect the final classification result, so the grid search mechanism can be applied.The following website may help.
http://www.csie.ntu.edu.tw/~cjlin/libsvm/#matlab
- Badges
- Science topic