what would be the right feature extraction methods for these images?

1 次查看（过去 30 天）

Hadeer tawfik 2016-10-5

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/305946-what-would-be-the-right-feature-extraction-methods-for-these-images

评论： Hadeer tawfik 2016-10-21

I am trying to classify cataract images, First i crop out the pupil area and save it in another folder, then i execute the wavelet transform on these cropped images using wavedec function and sym8 filter, then i take the approximation coefficients as the feature vector and the final step would be sending these feature vectors to neural network, after trying the neural network to classify my data set which contains 51 images,I have tried 10,20, 50, 70 hidden layers,But It shows that only 90% is the percentage of the correct classified images. so I want to increase this percentage,

So, any suggestion to use something other than wavelet?
why when i increased the dataset size to 83 images the neural network showed me bad results such as 70 or 60%?!
should i stop using Neural and start looking for another classifier?

6 个评论
显示 4更早的评论隐藏 4更早的评论

Greg Heath 2016-10-21

在 MATLAB Online 中打开

I assume you meant MATRIX instead of VECTOR. Even then, the matrix sizes for N pairs of I-dimensional "I"nput vectors and corresponding O-dimensional "O"utput target vectors are

 [ I N ] = size(input)
 [ O N ] = size(target)

For a 2-class classifier, O = 2, therefore I do not understand the value 5.

Using a NN input vector of length 8600 makes no sense at all.

Please explain your numbers w.r.t. a NN classifier with I-H-O node topology.

Confused,

Greg

Hadeer tawfik 2016-10-21

I am so sorry for this confusion, I have 83 images. after applying the wavelet using wavedec, I have a 87x87 matrix for every image that i changed it into a vector using reshape so the vector will be 1x7569 for each image.then i put all of the feature vectors into a 7569x83 matrix that will be fed to neural network , then i prepard the target matrix with is 5x83 which means that i want to classify the 83 image into 5 classes.I am so sorry for replaying late but it is due to the different time zone and sorry for not understanding your question earlier, Thank you so much

请先登录，再进行评论。

请先登录，再回答此问题。