KeCoDe_ENCODER_ACHROM applies DCT-based image coding algorithms to the input 256*256 gray scale image to obtain a file with the desired entropy (in bits/pixel). The program stores the encoded file in the folder\name provided by the user. Particular appendices to the given file name will be added depending on the selected algorithm (see names below). The encoded file can be decoded with KeCoDe_DECODER.M Besides, the encoder routine also gives the decoded image, the entropy of the code (in bits/pix) and a number of distortion measures. Algorithms 1-4 below use quantizers based on human vision models of increasing accuracy. Algorithms 5-7 are based on similar vision models and also use SVM based selection of transform coefficients. * MODEL 1: JPEG91. Wallace, Comm. of the ACM, Vol.34(4):30-44, (1991) block DCT transform coding with quantization based on JPEG-like linear CSF vision model. It does not include masking relations at all. The name of the encoded file will be: name_jpeg91_entropy_1.zip Note, though, that this is not exactly the JPEG standard since (1) 16*16 DCT blocks are used, (2) run length encoding, DPCM DC encoding and block information arrangement are done in a non-standard way, and (3) the final entropy coding is done by the Matlab zip routine. Nevertheless this implementation is the appropriate for fair comparison with the rest of the algorithms since the above (non-fundamental) details are implemented in the same way. * MODEL 2: Malo99. Malo et al., Electr. Lett., Vol.35(13):1067-1068, (1999) block DCT transform coding with quantization based on a simple (point-wise) non-linear masking model. It includes auto-masking but it does not include masking relations among coefficients. The name of the encoded file will be: name_malo99_entropy_1.zip * MODEL 3: Malo06. Malo et al. IEEE Trans. Im. Proc., Vol.15(1):68-80 (2006) block DCT transform plus non-linear divisive normalization transform and uniform quantization. This is the proper way to take frequency selectivity and all the masking relations into account in the quantization process. The name of the encoded file will be: name_malo06_entropy_1.zip * MODEL 4: Robinson03. Robinson & Kecman, IEEE Trans. Neur.Nets., Vol.14(4):950-958 (2003) block DCT transform plus CSF inspired constant insensitivity SVM coefficient selection (RKi-1). SVM based on a rough linear vision model. The name of the encoded file will be: name_robinson03_entropy_1.zip * MODEL 5: Gomez05. Gomez et al., IEEE Trans. Neur. Nets., Vol.16(6):1574-1581 (2005) block DCT transform coding plus CSF adaptive insensitivity SVM coefficient selection. SVM based on an accurate linear vision model. The name of the encoded file will be: name_gomez05_entropy_1.zip * MODEL 6: Camps08. Camps et al., J. Mach. Learn. Res., Vol.9(1):49-66 (2008) block DCT transform coding plus divisive normalization and constant insensitivity SVM coefficient selection. SVM trained in a vision model domain that takes into account frequency seletivity and masking relations among coefficients. The name of the encoded file will be: name_camps08_entropy_1.zip The bit rate of the above algorithms is controlled by different parameters: * Algorithms 1-3 depend on the Control Parameter, 'CP'. Smaller CP values imply more coarse quantization thus giving smaller files and more distorted images. * Algorithms 4-6 depend on two parameters: (1) the insensitivity parameter of the SVM, 'Epsilon'. (2) the number of bits used to encode the SVM weights, 'Bits'. For a fixed number of bits, smaller Epsilon values imply keeping more support vectors (or coefficients) and hence larger files and better quality images. The user has to provide a target entropy value. The program then sets the values of the control parameters (CP, or Epsilon and Bits) and iteratively modifies them to achieve the target entropy for the particular image. SYNTAX: [Results] = KeCoDe_encoder_achrom(Im,MODEL,'output_folder','name',target_entropies,Num_iterat) Input variables: ---------------- * Im : 256*256 image matrix double precision numbers in the range [0 255] * MODEL : 1-6 * 'output_folder': String with the folder where it will be written the output file(s) * 'name' : String with the name of the output file Note that an appendix to this name will be added depending on the coding algorithm. * target_entropies: Vector containing the set of target entropies (an image can be compressed at different entropies with a single call to this function) There will be as many output files as target entropy values. * Num_iterat : Number of iterations to look for the target entropy Output: ------- * Results : Struct variable with the following fields - Results(i).Image = Decoded image corresponding to the i-th value of the target entropy vector. - Results(i).Entropy = Entropy (in bits/pix): file_size/256^2 - Results(i).RMSE = RMSE distortion of the i-th decoded image - Results(i).SSIM = Structural SIMilarity Index of the i-th decoded image. (See Wang et al. IEEE Tr. Im. Proc., 2004 for a description of this distortion measure) - Results(i).MPE_linear= Maximum Perceptual Error of the i-th decoded image based on a linear CSF vision model. (See Gomez et al. IEEE Tr. Neur. Nets., 2005 for a description of this distortion measure) - Results(i).MPE_non_linear= Maximum Perceptual Error of the i-th decoded image based on a non linear vision model. (See Camps et al. JMLR, 2008 for a description of this distortion measure)
0001 0002 % 0003 % KeCoDe_ENCODER_ACHROM applies DCT-based image coding algorithms to the 0004 % input 256*256 gray scale image to obtain a file with the desired 0005 % entropy (in bits/pixel). 0006 % 0007 % The program stores the encoded file in the folder\name provided by the user. 0008 % Particular appendices to the given file name will be added depending on the 0009 % selected algorithm (see names below). 0010 % 0011 % The encoded file can be decoded with KeCoDe_DECODER.M 0012 % Besides, the encoder routine also gives the decoded image, the entropy 0013 % of the code (in bits/pix) and a number of distortion measures. 0014 % 0015 % Algorithms 1-4 below use quantizers based on human vision models of increasing 0016 % accuracy. Algorithms 5-7 are based on similar vision models and also use 0017 % SVM based selection of transform coefficients. 0018 % 0019 % * MODEL 1: JPEG91. Wallace, Comm. of the ACM, Vol.34(4):30-44, (1991) 0020 % block DCT transform coding with quantization based on JPEG-like 0021 % linear CSF vision model. It does not include masking relations at all. 0022 % The name of the encoded file will be: name_jpeg91_entropy_1.zip 0023 % 0024 % Note, though, that this is not exactly the JPEG standard since 0025 % (1) 16*16 DCT blocks are used, (2) run length encoding, DPCM DC encoding 0026 % and block information arrangement are done in a non-standard way, and (3) 0027 % the final entropy coding is done by the Matlab zip routine. 0028 % Nevertheless this implementation is the appropriate for fair 0029 % comparison with the rest of the algorithms since the above (non-fundamental) 0030 % details are implemented in the same way. 0031 % 0032 % * MODEL 2: Malo99. Malo et al., Electr. Lett., Vol.35(13):1067-1068, (1999) 0033 % block DCT transform coding with quantization based on a 0034 % simple (point-wise) non-linear masking model. It includes 0035 % auto-masking but it does not include masking relations among 0036 % coefficients. 0037 % The name of the encoded file will be: name_malo99_entropy_1.zip 0038 % 0039 % * MODEL 3: Malo06. Malo et al. IEEE Trans. Im. Proc., Vol.15(1):68-80 (2006) 0040 % block DCT transform plus non-linear divisive normalization 0041 % transform and uniform quantization. This is the proper way to take 0042 % frequency selectivity and all the masking relations into account in 0043 % the quantization process. 0044 % The name of the encoded file will be: name_malo06_entropy_1.zip 0045 % 0046 % * MODEL 4: Robinson03. Robinson & Kecman, IEEE Trans. Neur.Nets., Vol.14(4):950-958 (2003) 0047 % block DCT transform plus CSF inspired constant insensitivity 0048 % SVM coefficient selection (RKi-1). SVM based on a rough 0049 % linear vision model. 0050 % The name of the encoded file will be: name_robinson03_entropy_1.zip 0051 % 0052 % * MODEL 5: Gomez05. Gomez et al., IEEE Trans. Neur. Nets., Vol.16(6):1574-1581 (2005) 0053 % block DCT transform coding plus CSF adaptive insensitivity 0054 % SVM coefficient selection. SVM based on an accurate linear vision 0055 % model. 0056 % The name of the encoded file will be: name_gomez05_entropy_1.zip 0057 % 0058 % * MODEL 6: Camps08. Camps et al., J. Mach. Learn. Res., Vol.9(1):49-66 (2008) 0059 % block DCT transform coding plus divisive normalization and 0060 % constant insensitivity SVM coefficient selection. SVM trained in a 0061 % vision model domain that takes into account frequency seletivity and 0062 % masking relations among coefficients. 0063 % The name of the encoded file will be: name_camps08_entropy_1.zip 0064 % 0065 % The bit rate of the above algorithms is controlled by different 0066 % parameters: 0067 % 0068 % * Algorithms 1-3 depend on the Control Parameter, 'CP'. 0069 % Smaller CP values imply more coarse quantization thus giving smaller 0070 % files and more distorted images. 0071 % 0072 % * Algorithms 4-6 depend on two parameters: 0073 % (1) the insensitivity parameter of the SVM, 'Epsilon'. 0074 % (2) the number of bits used to encode the SVM weights, 'Bits'. 0075 % For a fixed number of bits, smaller Epsilon values imply keeping more 0076 % support vectors (or coefficients) and hence larger files and better 0077 % quality images. 0078 % 0079 % The user has to provide a target entropy value. The program then 0080 % sets the values of the control parameters (CP, or Epsilon and Bits) and 0081 % iteratively modifies them to achieve the target entropy for the particular image. 0082 % 0083 % SYNTAX: 0084 % [Results] = KeCoDe_encoder_achrom(Im,MODEL,'output_folder','name',target_entropies,Num_iterat) 0085 % 0086 % Input variables: 0087 % ---------------- 0088 % * Im : 256*256 image matrix double precision numbers in the range [0 255] 0089 % * MODEL : 1-6 0090 % * 'output_folder': String with the folder where it will be written the output file(s) 0091 % * 'name' : String with the name of the output file 0092 % Note that an appendix to this name will be added 0093 % depending on the coding algorithm. 0094 % * target_entropies: Vector containing the set of target entropies 0095 % (an image can be compressed at different entropies 0096 % with a single call to this function) 0097 % There will be as many output files as target entropy 0098 % values. 0099 % * Num_iterat : Number of iterations to look for the target entropy 0100 % 0101 % Output: 0102 % ------- 0103 % * Results : Struct variable with the following fields 0104 % 0105 % - Results(i).Image = Decoded image corresponding to the i-th value 0106 % of the target entropy vector. 0107 % - Results(i).Entropy = Entropy (in bits/pix): file_size/256^2 0108 % - Results(i).RMSE = RMSE distortion of the i-th decoded image 0109 % - Results(i).SSIM = Structural SIMilarity Index of the i-th 0110 % decoded image. 0111 % (See Wang et al. IEEE Tr. Im. Proc., 2004 0112 % for a description of this distortion measure) 0113 % - Results(i).MPE_linear= Maximum Perceptual Error of the i-th decoded 0114 % image based on a linear CSF vision model. 0115 % (See Gomez et al. IEEE Tr. Neur. Nets., 2005 0116 % for a description of this distortion measure) 0117 % - Results(i).MPE_non_linear= Maximum Perceptual Error of the i-th decoded 0118 % image based on a non linear vision model. 0119 % (See Camps et al. JMLR, 2008 for a description 0120 % of this distortion measure) 0121 % 0122 function [Results] = KeCoDe_encoder_achrom(Im,algorit,directorio,ficherin,entropias,N_it) 0123 warning('off','MATLAB:dispatcher:InexactMatch') 0124 Im=double(Im); 0125 if algorit==1 0126 algoritmo=6; 0127 0128 Desired_entropy=[0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 1 1.2 1.4 1.6 1.8 2 2.2]; 0129 Approximate_uCP=[1.5 2.6 3.3 3.8 4.4 4.9 5.3 5.7 6.5 7.3 8.1 9 9.9 11 12.2]; 0130 for i=1:length(entropias) 0131 if entropias(i)>max(Desired_entropy), 0132 parametro(i)=12.5; 0133 elseif entropias(i)<min(Desired_entropy), 0134 parametro(i)=1.2; 0135 else 0136 parametro(i)=interp1(Desired_entropy,Approximate_uCP,entropias(i)); 0137 end 0138 end 0139 elseif algorit==2 0140 algoritmo=7; 0141 0142 Desired_entropy=[ 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 1 1.2 1.4 1.6 1.8 2 2.2]; 0143 Approximate_uCP=[ 1.3 2.3 3 3.3 3.7 4.1 4.5 4.8 5.3 5.8 6.3 6.8 7.4 7.8 8.3]; 0144 for i=1:length(entropias) 0145 if entropias(i)>max(Desired_entropy), 0146 parametro(i)=8.4; 0147 elseif entropias(i)<min(Desired_entropy), 0148 parametro(i)=1.1; 0149 else 0150 parametro(i)=interp1(Desired_entropy,Approximate_uCP,entropias(i)); 0151 end 0152 end 0153 elseif algorit==3 0154 algoritmo=9; 0155 0156 Desired_entropy=[ 0.2 0.3 0.4 0.5 0.6 0.7 0.8 1 1.2 1.4 1.6 1.8 2 2.2]; 0157 Approximate_uCP=[ 0.3 1.0 1.7 2.1 2.6 2.9 3.3 3.9 4.4 4.9 5.3 5.6 6 6.3]; 0158 for i=1:length(entropias) 0159 if entropias(i)>max(Desired_entropy), 0160 parametro(i)=6.5; 0161 elseif entropias(i)<min(Desired_entropy), 0162 parametro(i)=0.15; 0163 else 0164 parametro(i)=interp1(Desired_entropy,Approximate_uCP,entropias(i)); 0165 end 0166 end 0167 elseif algorit==4 0168 algoritmo=2; 0169 0170 Desired_entropy=[ 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 1 1.2 1.4 1.6 1.8 2 2.2]; 0171 Approx_Epsilon=[ 23.6 2.9 0.46 0.69 0.012 0.037 0.002 0.016 0.013 0.0004 0.0034 0.0053 0.0016 0.0037 0.0017]; 0172 for i=1:length(entropias) 0173 if entropias(i)>max(Desired_entropy), 0174 parametro(i)=0.001; 0175 elseif entropias(i)<min(Desired_entropy), 0176 parametro(i)=25; 0177 else 0178 parametro(i)=interp1(Desired_entropy,Approx_Epsilon,entropias(i)); 0179 end 0180 end 0181 elseif algorit==5 0182 algoritmo=3; 0183 0184 Desired_entropy=[0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 1 1.2 1.4 1.6 1.8 2 2.2]; 0185 Approx_Epsilon=[15.7 0.18 0.10 0.065 0.009 0.026 0.002 0.012 0.011 0.0003 0.003 0.005 0.0015 0.0033 0.0016]; 0186 for i=1:length(entropias) 0187 if entropias(i)>max(Desired_entropy), 0188 parametro(i)=0.001; 0189 elseif entropias(i)<min(Desired_entropy), 0190 parametro(i)=18; 0191 else 0192 parametro(i)=interp1(Desired_entropy,Approx_Epsilon,entropias(i)); 0193 end 0194 end 0195 elseif algorit==6 0196 algoritmo=4; 0197 0198 Desired_entropy=[0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 1 1.2 1.4 1.6 1.8 2 2.2]; 0199 Approx_Epsilon=[0.79 0.36 0.17 0.08 0.025 0.025 0.004 0.01 0.004 0.0027 0.0024 0.003 0.0036 0.0002 0.0012]; 0200 for i=1:length(entropias) 0201 if entropias(i)>max(Desired_entropy), 0202 parametro(i)=0.001; 0203 elseif entropias(i)<min(Desired_entropy), 0204 parametro(i)=1; 0205 else 0206 parametro(i)=interp1(Desired_entropy,Approx_Epsilon,entropias(i)); 0207 end 0208 end 0209 else 0210 disp('Not a valid algorithm selection') 0211 algoritmo=1000; 0212 end 0213 if algoritmo<100 0214 0215 [perfil,K,exponente] = computing_parameters_entropy(algoritmo); 0216 if(algoritmo <= 4) 0217 Results = entropy_svr(algoritmo,entropias,Im,parametro,perfil,K,exponente,directorio,ficherin,N_it); 0218 else 0219 Results = entropy_ucp(algoritmo,entropias,Im,parametro,exponente,directorio,ficherin,N_it); 0220 end 0221 end