Restoration of Partial Blurred Image Based on Blur Detection and Classification

Algorithm 2

Blur classification algorithm.

Input: sample images (), blurred image

Output: blur class of

 Step Feature Extraction

 ;   % Gradient feature extraction

 FT = (1 + fftshift(abs(Fourier()))); % Fourier transformation

  = Radon(FT);  % Radon transformation feature

 edge = Canny(FT);    % Compute the edge of FT

  = Houghline(edge);  % Extract the line feature

  = Houghcirle(edge):  % Extract the circle feature

 Feature normalization;

 Step SVM Traning

 SVM₁ = Svmtrain(, ); % SVM₁ Training

 SVM₂ = Svmtrain(, );    % SVM₂ Training

 Step Classification

 Feature Extraction of y as Step ;

 defocus = SVM₁(, ); % SVM₁ prediction

  If defocus = 1

   c = defocus;

   exit;

  Else

   motion = SVM₂(, ); % SVM₂ prediction

   If motion = 1

    c = motion;

    exit;

   Else

    c = blend;

    exit;

   End If

  End If