MATLAB实现SIFT特征提取算法

# SIFT-on-MATLAB An implementation of [Distinctive image features from scale-invariant keypoints](https://scholar.google.com/citations?view_op=view_citation&hl=en&user=8vs5HGYAAAAJ&citation_for_view=8vs5HGYAAAAJ:u_35RYKgDlwC), created by [David Lowe](https://scholar.google.com/citations?user=8vs5HGYAAAAJ). This is a term project for "Advanced Topics in Medical Image Analysis" course at Middle East Technical University. # Usage Read an image in MatLab and convert it into gray scale image then use it as input for `SIFT` function. - `SIFT(Image, Octaves, Scales, Sigma):` Main function takes **gray scale image**, **number of octaves**, **number of scales per octaves** and **initial value for sigma**. Returns key points inside a cell array. Example: ``` image = imread('image.jpg'); image = double(rgb2gray(image)); keyPoints = SIFT(image,3,5,1.6); ``` Key points created in the process are objects. All of the key points are returned in a cell array. Each key point contains: - `coordinates():`returns **[x, y]** coordinate of the key point on image. - `direction():`Returns general Direction of the key point. - `magnitude():`Returns the magnitude of general direction vector. - `octave():`Returns number of the octave which the key point extracted from. - `scale():` Returns sigma value which the image is convolved with. - `descriptor():`Returns a vector containing the descriptor. # Image Visualizer This function's main purpose is to illustrate the keypoints on the image. The function `SIFTKeypointVisualizer` will be called after extracting keypoints. Example: ``` image = SIFTKeypointVisualizer(image,keyPoints); imshow(uint8(image)); ``` # Dependencies This code is dependent on some functions within `Image Processing Toolbox` version `9.2`. This code should be compatible with any MatLab newer than `R2015a`. # Disclaimer All the rights to **"Distinctive image features from scale-invariant keypoints"** are reserved for **[University of British Columbia](http://www.cs.ubc.ca/~lowe/keypoints/)**. *A license must be obtained from the University of British Columbia for any commercial applications.* # Contact For further information please contact [zabardast.amin@metu.edu.tr](mailto:zabardast.amin@metu.edu.tr) or [kilic.ozkan@metu.edu.tr](mailto:kilic.ozkan@metu.edu.tr).