Picture comparison, latest.

I have completed running C++ code for image stitching using ASIFT.

Which robust, reliable, SOTA, industrious, accurate, fast, contemporary image recognition solution you can propose?

What is the function(methode) to see if 2 images have similar vectors?

Answer 1

[Rahul]

Hi @Pavl M.,

In order to apply image recognition on the image obtained after image stitching using ASIFT, consider using 'Image Recognition' with MATLAB, which provides features for image labeling and exploring deep learning and machine learning algorithms for object recognition.

In order to compare images for similarity vectors, consider one approach of using the 'detectSURFFeatures', 'extractFeatures', and 'matchFeatures' functions to find and match features between two images. Here is an example:

% Detect features
points1 = detectSURFFeatures(image1);
points2 = detectSURFFeatures(image2);

% Extract features
[features1, validPoints1] = extractFeatures(image1, points1);
[features2, validPoints2] = extractFeatures(image2, points2);

% Match features
indexPairs = matchFeatures(features1, features2);

Other approaches which can be used for comparing images for similarity:

Histogram Comparision - 'imhist' and 'corr2' functions can be used. Here is an example

hist1 = imhist(image1);
hist2 = imhist(image2);
similarity = corr2(hist1, hist2); 

Structural Similarity Index (SSIM) = 'ssim' function can be used. Here is an example

similarity = ssim(image1, image2);

Refer to the following MathWorks documentations to know more:

‘Image Recognition with MATLAB’: https://www.mathworks.com/discovery/image-recognition-matlab.html#image-recognition-with-matlab

'detectSURFFeatures': https://www.mathworks.com/help/vision/ref/detectsurffeatures.html

'extractFeatures': https://www.mathworks.com/help/vision/ref/extractfeatures.html

'matchFeatures': https://www.mathworks.com/help/vision/ref/matchfeatures.html

'imhist': https://www.mathworks.com/help/images/ref/imhist.html

'ssim': https://www.mathworks.com/help/images/ref/ssim.html

Hope this helps! Thanks.