Features extraction to identify lesions kind among many in the database matlab

what kind of COLOR and TEXTURE based treatment and command should i use to get the the information i'd want regarding every pic in the database and match it with one taken from a user,kindly note than many pics will be used for each skin issue (features extraction),i'd also wanna know where i can find some high quality skin issues pics(website name or a known dataset).

[John BG]

do you think it would be possible for you to post a picture and tell us what do you want to do?

like spotting marked cells, measuring distance, fractures ..

Answer 1

[Image Analyst]

Start your search with Google: https://www.google.com/?gws_rd=ssl#q=cbir

[Elias Unk]

Lol,i got past that phase i found some pretty interesting previous IEEE published projects indicating the extract four statistics (mean, standard deviation, variation and skewness) from segmented lesion regions over individual channels of six different color spaces : RGB, HSV, YCbCr, NTSc, CIE L*u*v and CIE L*a*b.And for texture extraction originally proposed a total of 14 features namely, Angular second moment, contrast, correlation, sum of variance, inverse difference moment, sum average, sum variance, sum entropy, entropy, difference variance, difference entropy, information measures of correlation, and maximal correlation coefficient. In order to obtain texture features 9, the normalized GLCM was computed for each of the four orientations ({00, 450, 900,1350}). That being said,i'm a starter in this domain so would've wanted some better clearer steps instead of sarcastic answers @Image Analyst.

[Image Analyst]

How was I supposed to know that you had already known about all those different things you could measure? There's a lot in your comment that you didn't share in your original question - different features, etc. I didn't even know that you knew about CBIR methods yet - most people haven't even heard of CBIR. Evidently you know more than I had assumed. You've even computed GLCM, moments, etc. You've probably also searched PubMed and VisionBib for skin cancer articles, so sorry but I have nothing to offer beyond those papers you've read. Good luck though.

Regarding finding sample medical images, you might try http://www.osirix-viewer.com/resources/dicom-image-library/ or else ask the sponsor of your research - whoever hired you to solve this problem.

[Elias Unk]

Will check them out..thanks for the help mate.