I have been working on extracting features from wholelside images in MATLAB. In that, i have so far succeeded in extracting 118 features from 92 training images. I want to select the feature that provides best two-class separability, the two classes being 'Infected' and 'Normal'. I evaluated the ks-density plot of each of the feature but couldn't get a numerical measure of the amount of overlap between the pdfs of a given feature from 'Normal' class and 'Infected' class.
Further, I tried ranking the features based on the method given in this paper: http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0082710#pone.0082710-Newcombe1.
But this method seems to point the best feature as something which has considerable amount of overlap of pdfs between the two classes. When I test the accuracy of the method by measuring the Bhattacharya distance between the feature vector of the two classes, it shows some other set of features that are best.
I am not able to decide which method to use for feature selection. Can someone give me some guidance?
My goal is to correlate the amount of overlap between the pdfs of the feature vector of the two classes with the results of a proper feature selection algorithm. In other words, what the chosen algorithm points out as the best feature for classification should have the least pdf overlap.<\p>
If also someone can give me guidance on how I can estimate the amount of overlap between a given feature vector of class 'Normal' and the same feature vector of class 'Infected', that would be great.
Related
I am working on a small project in Matlab just because of my interest in image processing and I have not studied a degree or a course related to image processing.
I want to understand a small concept about feature extraction and feature vectors. I have read some articles about that and in general I can understand that, but my question is:
For example, I want to extract some information from different objects of a binary image, the information is about length, width and distance between the objects. In one application I want to extract the features on which I want to apply some algorithms to compute width of all the objects and ignore the length and distance. Can we name this as feature extraction regarding the width? And storing them in different vectors as Feature Vectors?
It makes me think that, I might be complicating the simple things. Should I use some other terminologies for this instead of feature extraction and feature vectors?
Please suggest me if I am going in the right direction or not?
Thank you!
Feature extraction is the process of computing numerical values on regions/objects/shapes/blobs detected in an image. [Sometimes the detection itself can be called extraction and the features need not be numbers.]
The feature values can indeed be stored in vectors, usually they fill a table. Sometimes they are structured in a more complicated way (such as a graph f.i.). Most of the time they are used for classification/recognition purposes or they can just be the output of the process on hand.
so I have a dataset of 77 patients cancer patients and 12500+ attributes. I have applied Principal Component Analysis in order to filter the attributes to only retain the ones the explain the most variance.
My question is, are there techniques in Matlab, other than PCA, to identify the attributes with the most predictive power?
There are two main ways to cleverly "reduce the dimensionality" of your dataset. One is Feature Transformation (that includes, for example, PCA), and the other one is Feature Selection.
It seems that you are looking for a Feature Selection algorithm, that would retain the most informative original attributes. On the contrary, a Feature Transformation algorithm will generate a new set of attributes!
As for your exact question, there are multiple choices you can make. Keep in mind that, naively, each Feature Selection algorithm will have to choose the best features according to "how well" those features alone can model the problem.
For a MATLAB built-in implementation, if you have the Statistics and Machine Learning Toolbox installed, you can use the "Sequential feature selection" function sequentialfs.
I have run LDA with MATLAB using the fitcdiscr function and predict.
I have a feeling there may be some bugs in my code however and as a sanity check would like to identify which features are being most heavily weighted in the classification.
Can this be done?
There is a Coeffs field in your fitted object containing all the relevant information http://uk.mathworks.com/help/stats/classificationdiscriminant-class.html
In particular, if you fit a linear LDA there will be Linear field which is the linear operator used for projection. However, one should bear in mind that value of coefficients of linear models are not feature importances. There is much more in that to consider. Weight can be big because your feature have small values or because there is a highly biased distribution of the values. If you need feature selection technique - use feature selection methods (like L1 regularized models) otherwise you might easily get wrong conclusions from your data.
I'm fairly new to MATLAB, but have acquainted myself with Simulink and Computer Vision over the past few days. My problem statement involves taking a traffic/highway video input and detecting if an accident has occurred.
I plan to do this by extracting the values of centroid to plot trajectory, velocity difference (between frames) and distance between two vehicles. I can successfully track the centroids, and aim to derive the rest of the features.
What I don't know is how to map these to ANN. I mean, every image has more than one vehicle blobs, which means, there are multiple centroids in a single frame/image. So, how does NN act on multiple inputs (the extracted features per vehicle) simultaneously? I am obviously missing the link. Help me figure it out please.
Also, am I looking at time series data?
I am not exactly sure about your question. The problem can be both time series data and not. You might be able to transform the time series version of the problem, such that it can be solved using ANN, but it is sort of a Maslow's hammer :). Also, Could you rephrase the problem.
As you said, you could give it features from two or three frames and then use the classifier to detect accident or not, but it might be difficult to train such a classifier. The problem is really difficult and the so you might need tons of training samples to get it right, esp really good negative samples (for examples cars travelling close to each other) etc.
There are multiple ways you can try to solve this problem of accident detection. For example : Build a classifier (ANN/SVM etc) to detect accidents without time series data. In which case your input would be accident images and non accident images or some sort of positive and negative samples for training and later images for test. In this specific case, you are not looking at the time series data. But here you might need lots of features to detect the same (this in some sense a single frame version of the problem).
The second method would be to use time series data, in which case you will have to detect the features, track the features (say using Lucas Kanade/Horn and Schunck) and then use the information about velocity and centroid to detect the accident. You might even be able to formulate it for HMMs.
My BE final year project is about sign language recognition. I'm terribly confused in choosing the right classification technique for patterns seen in the video of signs generated by a dumb user. I learned neural nets(NN) are better than hidden markov model in several aspects but fine tuning the parameters of NN requires a lot of time. Further, some reports say that Support Vector Machine are better in performance than NN. What do I choose among these alternatives or are there any other better alternatives so that it would be feasible to complete my project within 4-5 months and I could continue with that field in my masters?
Actually the system will be fed with real time video and we intend to recognize the hand postures and spatiotemporal gestures. So, its the entire sentences I'm trying to find.
On the basis of studies till now, I'm making my mind to use
1. Hu moments & eigenspace size functions to represent hand shapes
2. SVM for posture classification &
3. Threshold HMM for spatiotemporal gesture recognition.
What would u comment in these decisions?