Revision history [back]

I am working on Face recognition as a part of my project. I am using the fisherfaces program provided in OpenCV website for the same. But the program code provided there does not have verification and the recognition always results in a subject in the database even though it is a false subject. The fisherfaces reconstruction does not perform well. The Similarity error for both the true test inputs (true test input - picked from the database) and false test inputs (false test input - NOT picked from the database) are quite close to each other and some even overlap with the range of values from true test inputs.

I performed an experiment on AT & T database which has 40 subjects.

I selected 20 out of them for database and rest 20 i used to provide false test input where the ideal behavior of false test input is not to get recognized to one of the images in database. But the normalized euclidean distance of the false test inputs are quite near and some even overlap with the normalized euclidean distance of the true test inputs. Please let me know if there is any method to set an appropriate threshold and hence make the algorithm work well.

Results for false_test_inputs

Identity: 2. Similarity error: 0.434669 Identity: 1. Similarity error: 0.440239 Identity: 3. Similarity error: 0.454631 Identity: 3. Similarity error: 0.428741 Identity: 3. Similarity error: 0.458096 Identity: 3. Similarity error: 0.463923 Identity: 3. Similarity error: 0.434852 Identity: 3. Similarity error: 0.439271 Identity: 5. Similarity error: 0.452804 Identity: 2. Similarity error: 0.441438 Identity: 0. Similarity error: 0.334382 Identity: 0. Similarity error: 0.313689 Identity: 0. Similarity error: 0.367018 Identity: 7. Similarity error: 0.380997 Identity: 0. Similarity error: 0.367258 Identity: 0. Similarity error: 0.357048 Identity: 0. Similarity error: 0.364141 Identity: 0. Similarity error: 0.368228 Identity: 0. Similarity error: 0.322856 Identity: 0. Similarity error: 0.298744 Identity: 0. Similarity error: 0.3978 Identity: 0. Similarity error: 0.354158 Identity: 8. Similarity error: 0.417423 Identity: 0. Similarity error: 0.42364 Identity: 0. Similarity error: 0.420527 Identity: 0. Similarity error: 0.361855 Identity: 0. Similarity error: 0.367132 Identity: 0. Similarity error: 0.373914 Identity: 0. Similarity error: 0.370478 Identity: 0. Similarity error: 0.343055 Identity: 3. Similarity error: 0.545497 Identity: 2. Similarity error: 0.509753 Identity: 3. Similarity error: 0.537281 Identity: 3. Similarity error: 0.533134 Identity: 8. Similarity error: 0.530642 Identity: 2. Similarity error: 0.572951 Identity: 3. Similarity error: 0.461317 Identity: 3. Similarity error: 0.492737 Identity: 3. Similarity error: 0.477977 Identity: 5. Similarity error: 0.46319 Identity: 3. Similarity error: 0.333414 Identity: 2. Similarity error: 0.301396 Identity: 2. Similarity error: 0.295346 Identity: 3. Similarity error: 0.327051 Identity: 3. Similarity error: 0.332282 Identity: 5. Similarity error: 0.299717 Identity: 3. Similarity error: 0.342065 Identity: 5. Similarity error: 0.304572 Identity: 3. Similarity error: 0.350087 Identity: 2. Similarity error: 0.311405 Identity: 3. Similarity error: 0.47027 Identity: 2. Similarity error: 0.327899 Identity: 8. Similarity error: 0.352923 Identity: 8. Similarity error: 0.347418 Identity: 8. Similarity error: 0.529454 Identity: 8. Similarity error: 0.522772 Identity: 3. Similarity error: 0.420531 Identity: 8. Similarity error: 0.362175 Identity: 8. Similarity error: 0.36085 Identity: 8. Similarity error: 0.362896 Identity: 5. Similarity error: 0.359937 Identity: 5. Similarity error: 0.375324 Identity: 5. Similarity error: 0.355757 Identity: 2. Similarity error: 0.290591 Identity: 2. Similarity error: 0.294904 Identity: 5. Similarity error: 0.357246 Identity: 1. Similarity error: 0.358283 Identity: 5. Similarity error: 0.370913 Identity: 2. Similarity error: 0.3373 Identity: 3. Similarity error: 0.359347 Identity: 0. Similarity error: 0.375202 Identity: 0. Similarity error: 0.318169 Identity: 0. Similarity error: 0.361512 Identity: 0. Similarity error: 0.357798 Identity: 0. Similarity error: 0.37323 Identity: 0. Similarity error: 0.327084 Identity: 0. Similarity error: 0.333594 Identity: 0. Similarity error: 0.350686 Identity: 0. Similarity error: 0.351189 Identity: 0. Similarity error: 0.327009 Identity: 3. Similarity error: 0.416531 Identity: 7. Similarity error: 0.412658 Identity: 3. Similarity error: 0.432793 Identity: 8. Similarity error: 0.458255 Identity: 8. Similarity error: 0.408839 Identity: 8. Similarity error: 0.406158 Identity: 8. Similarity error: 0.457394 Identity: 8. Similarity error: 0.426322 Identity: 8. Similarity error: 0.415844 Identity: 9. Similarity error: 0.34565 Identity: 5. Similarity error: 0.408107 Identity: 5. Similarity error: 0.41425 Identity: 5. Similarity error: 0.407549 Identity: 5. Similarity error: 0.3877 Identity: 1. Similarity error: 0.396434 Identity: 7. Similarity error: 0.404262 Identity: 5. Similarity error: 0.413639 Identity: 5. Similarity error: 0.383368 Identity: 7. Similarity error: 0.408296 Identity: 7. Similarity error: 0.395991

Results for true_test_inputs

Identity: 8. Similarity error: 0.291803 Identity: 8. Similarity error: 0.381701 Identity: 8. Similarity error: 0.455005 Identity: 8. Similarity error: 0.378799 Identity: 8. Similarity error: 0.480244 Identity: 8. Similarity error: 0.456196 Identity: 8. Similarity error: 0.419657 Identity: 8. Similarity error: 0.361412 Identity: 8. Similarity error: 0.43065 Identity: 8. Similarity error: 0.470874 Identity: 3. Similarity error: 0.315804 Identity: 3. Similarity error: 0.321898 Identity: 3. Similarity error: 0.289971 Identity: 3. Similarity error: 0.307756 Identity: 3. Similarity error: 0.297371 Identity: 3. Similarity error: 0.29975 Identity: 3. Similarity error: 0.304808 Identity: 3. Similarity error: 0.345169 Identity: 3. Similarity error: 0.319984 Identity: 3. Similarity error: 0.34902 Identity: 7. Similarity error: 0.263428 Identity: 7. Similarity error: 0.27732 Identity: 7. Similarity error: 0.288591 Identity: 7. Similarity error: 0.277034 Identity: 7. Similarity error: 0.296871 Identity: 7. Similarity error: 0.335843 Identity: 7. Similarity error: 0.224683 Identity: 7. Similarity error: 0.252968 Identity: 7. Similarity error: 0.246976 Identity: 7. Similarity error: 0.253152 Identity: 5. Similarity error: 0.286976 Identity: 5. Similarity error: 0.306188 Identity: 5. Similarity error: 0.313551 Identity: 5. Similarity error: 0.305116 Identity: 5. Similarity error: 0.261896 Identity: 5. Similarity error: 0.300823 Identity: 5. Similarity error: 0.295371 Identity: 5. Similarity error: 0.258079 Identity: 5. Similarity error: 0.300896 Identity: 5. Similarity error: 0.303808 Identity: 0. Similarity error: 0.321324 Identity: 0. Similarity error: 0.282321 Identity: 0. Similarity error: 0.293295 Identity: 0. Similarity error: 0.27549 Identity: 0. Similarity error: 0.332434 Identity: 0. Similarity error: 0.292148 Identity: 0. Similarity error: 0.285137 Identity: 0. Similarity error: 0.327177 Identity: 0. Similarity error: 0.27173 Identity: 0. Similarity error: 0.254691 Identity: 4. Similarity error: 0.323429 Identity: 4. Similarity error: 0.396412 Identity: 4. Similarity error: 0.335021 Identity: 4. Similarity error: 0.307829 Identity: 4. Similarity error: 0.401246 Identity: 4. Similarity error: 0.359469 Identity: 4. Similarity error: 0.368059 Identity: 4. Similarity error: 0.380822 Identity: 4. Similarity error: 0.394788 Identity: 4. Similarity error: 0.433584 Identity: 1. Similarity error: 0.295633 Identity: 1. Similarity error: 0.302817 Identity: 1. Similarity error: 0.310627 Identity: 1. Similarity error: 0.287672 Identity: 1. Similarity error: 0.306451 Identity: 1. Similarity error: 0.322189 Identity: 1. Similarity error: 0.277154 Identity: 1. Similarity error: 0.283235 Identity: 1. Similarity error: 0.294519 Identity: 1. Similarity error: 0.304006 Identity: 9. Similarity error: 0.346696 Identity: 9. Similarity error: 0.297305 Identity: 9. Similarity error: 0.365462 Identity: 9. Similarity error: 0.301169 Identity: 9. Similarity error: 0.334413 Identity: 9. Similarity error: 0.347344 Identity: 9. Similarity error: 0.323791 Identity: 9. Similarity error: 0.348167 Identity: 9. Similarity error: 0.319981 Identity: 9. Similarity error: 0.355038 Identity: 2. Similarity error: 0.307948 Identity: 2. Similarity error: 0.31272 Identity: 2. Similarity error: 0.293307 Identity: 2. Similarity error: 0.30623 Identity: 2. Similarity error: 0.266553 Identity: 2. Similarity error: 0.319812 Identity: 2. Similarity error: 0.285 Identity: 2. Similarity error: 0.300466 Identity: 2. Similarity error: 0.258334 Identity: 2. Similarity error: 0.274881 Identity: 6. Similarity error: 0.465977 Identity: 6. Similarity error: 0.338479 Identity: 6. Similarity error: 0.407876 Identity: 6. Similarity error: 0.38913 Identity: 6. Similarity error: 0.366433 Identity: 6. Similarity error: 0.361704 Identity: 6. Similarity error: 0.424361 Identity: 6. Similarity error: 0.446707 Identity: 6. Similarity error: 0.414173 Identity: 6. Similarity error: 0.409185

I am working on Face recognition as a part of my project. I am using the fisherfaces program provided in OpenCV website for the same. But the program code provided there does not have verification and the recognition always results in a subject in the database even though it is a false subject. The fisherfaces reconstruction does not perform well. The Similarity error for both the true test inputs (true test input - picked from the database) and false test inputs (false test input - NOT picked from the database) are quite close to each other and some even overlap with the range of values from true test inputs.

I performed an experiment on AT & T database which has 40 subjects.

I selected 20 out of them for database and rest 20 i used to provide false test input where the ideal behavior of false test input is not to get recognized to one of the images in database. But the normalized euclidean distance of the false test inputs are quite near and some even overlap with the normalized euclidean distance of the true test inputs. Please let me know if there is any method to set an appropriate threshold and hence make the algorithm work well.

Results for false_test_inputs

Identity: 2. Similarity error: 0.434669 Identity: 1. Similarity error: 0.440239 Identity: 3. Similarity error: 0.454631 Identity: 3. Similarity error: 0.428741 Identity: 3. Similarity error: 0.458096 Identity: 3. Similarity error: 0.463923 Identity: 3. Similarity error: 0.434852 Identity: 3. Similarity error: 0.439271 Identity: 5. Similarity error: 0.452804 Identity: 2. Similarity error: 0.441438 Identity: 0. Similarity error: 0.334382 Identity: 0. Similarity error: 0.313689 Identity: 0. Similarity error: 0.367018 Identity: 7. Similarity error: 0.380997 Identity: 0. Similarity error: 0.367258 Identity: 0. Similarity error: 0.357048 Identity: 0. Similarity error: 0.364141 Identity: 0. Similarity error: 0.368228 Identity: 0. Similarity error: 0.322856 Identity: 0. Similarity error: 0.298744 Identity: 0. Similarity error: 0.3978 Identity: 0. Similarity error: 0.354158 Identity: 8. Similarity error: 0.417423 Identity: 0. Similarity error: 0.42364 Identity: 0. Similarity error: 0.420527 Identity: 0. Similarity error: 0.361855 Identity: 0. Similarity error: 0.367132 Identity: 0. Similarity error: 0.373914 Identity: 0. Similarity error: 0.370478 Identity: 0. Similarity error: 0.343055 Identity: 3. Similarity error: 0.545497 Identity: 2. Similarity error: 0.509753 Identity: 3. Similarity error: 0.537281 Identity: 3. Similarity error: 0.533134 Identity: 8. Similarity error: 0.530642 Identity: 2. Similarity error: 0.572951 Identity: 3. Similarity error: 0.461317 Identity: 3. Similarity error: 0.492737 Identity: 3. Similarity error: 0.477977 Identity: 5. Similarity error: 0.46319 Identity: 3. Similarity error: 0.333414 Identity: 2. Similarity error: 0.301396 Identity: 2. Similarity error: 0.295346 Identity: 3. Similarity error: 0.327051 Identity: 3. Similarity error: 0.332282 Identity: 5. Similarity error: 0.299717 Identity: 3. Similarity error: 0.342065 Identity: 5. Similarity error: 0.304572 Identity: 3. Similarity error: 0.350087 Identity: 2. Similarity error: 0.311405 Identity: 3. Similarity error: 0.47027 Identity: 2. Similarity error: 0.327899 Identity: 8. Similarity error: 0.352923 Identity: 8. Similarity error: 0.347418 Identity: 8. Similarity error: 0.529454 Identity: 8. Similarity error: 0.522772 Identity: 3. Similarity error: 0.420531 Identity: 8. Similarity error: 0.362175 Identity: 8. Similarity error: 0.36085 Identity: 8. Similarity error: 0.362896 Identity: 5. Similarity error: 0.359937 Identity: 5. Similarity error: 0.375324 Identity: 5. Similarity error: 0.355757 Identity: 2. Similarity error: 0.290591 Identity: 2. Similarity error: 0.294904 Identity: 5. Similarity error: 0.357246 Identity: 1. Similarity error: 0.358283 Identity: 5. Similarity error: 0.370913 Identity: 2. Similarity error: 0.3373 Identity: 3. Similarity error: 0.359347 Identity: 0. Similarity error: 0.375202 Identity: 0. Similarity error: 0.318169 Identity: 0. Similarity error: 0.361512 Identity: 0. Similarity error: 0.357798 Identity: 0. Similarity error: 0.37323 Identity: 0. Similarity error: 0.327084 Identity: 0. Similarity error: 0.333594 Identity: 0. Similarity error: 0.350686 Identity: 0. Similarity error: 0.351189 Identity: 0. Similarity error: 0.327009 Identity: 3. Similarity error: 0.416531 Identity: 7. Similarity error: 0.412658 Identity: 3. Similarity error: 0.432793 Identity: 8. Similarity error: 0.458255 Identity: 8. Similarity error: 0.408839 Identity: 8. Similarity error: 0.406158 Identity: 8. Similarity error: 0.457394 Identity: 8. Similarity error: 0.426322 Identity: 8. Similarity error: 0.415844 Identity: 9. Similarity error: 0.34565 Identity: 5. Similarity error: 0.408107 Identity: 5. Similarity error: 0.41425 Identity: 5. Similarity error: 0.407549 Identity: 5. Similarity error: 0.3877 Identity: 1. Similarity error: 0.396434 Identity: 7. Similarity error: 0.404262 Identity: 5. Similarity error: 0.413639 Identity: 5. Similarity error: 0.383368 Identity: 7. Similarity error: 0.408296 Identity: 7. Similarity error: 0.3959910.42364

Results for true_test_inputs

Identity: 8. Similarity error: 0.291803 Identity: 8. Similarity error: 0.381701 Identity: 8. Similarity error: 0.455005 Identity: 8. Similarity error: 0.378799 Identity: 8. Similarity error: 0.480244 Identity: 8. Similarity error: 0.456196 Identity: 8. Similarity error: 0.419657 Identity: 8. Similarity error: 0.361412 Identity: 8. Similarity error: 0.43065 Identity: 8. Similarity error: 0.470874 Identity: 3. Similarity error: 0.315804 Identity: 3. Similarity error: 0.321898 Identity: 3. Similarity error: 0.289971 Identity: 3. Similarity error: 0.307756 Identity: 3. Similarity error: 0.297371 Identity: 3. Similarity error: 0.29975 Identity: 3. Similarity error: 0.304808 Identity: 3. Similarity error: 0.345169 Identity: 3. Similarity error: 0.319984 Identity: 3. Similarity error: 0.34902 Identity: 7. Similarity error: 0.263428 Identity: 7. Similarity error: 0.27732 Identity: 7. Similarity error: 0.288591 Identity: 7. Similarity error: 0.277034 Identity: 7. Similarity error: 0.296871 Identity: 7. Similarity error: 0.335843 Identity: 7. Similarity error: 0.224683 Identity: 7. Similarity error: 0.252968 Identity: 7. Similarity error: 0.246976 Identity: 7. Similarity error: 0.253152 Identity: 5. Similarity error: 0.286976 Identity: 5. Similarity error: 0.306188 Identity: 5. Similarity error: 0.313551 Identity: 5. Similarity error: 0.305116 Identity: 5. Similarity error: 0.261896 Identity: 5. Similarity error: 0.300823 Identity: 5. Similarity error: 0.295371 Identity: 5. Similarity error: 0.258079 Identity: 5. Similarity error: 0.300896 Identity: 5. Similarity error: 0.303808 Identity: 0. Similarity error: 0.321324 Identity: 0. Similarity error: 0.282321 Identity: 0. Similarity error: 0.293295 Identity: 0. Similarity error: 0.27549 Identity: 0. Similarity error: 0.332434 Identity: 0. Similarity error: 0.292148 Identity: 0. Similarity error: 0.285137 Identity: 0. Similarity error: 0.327177 Identity: 0. Similarity error: 0.27173 Identity: 0. Similarity error: 0.254691 Identity: 4. Similarity error: 0.323429 Identity: 4. Similarity error: 0.396412 Identity: 4. Similarity error: 0.335021 Identity: 4. Similarity error: 0.307829 Identity: 4. Similarity error: 0.401246 Identity: 4. Similarity error: 0.359469 Identity: 4. Similarity error: 0.368059 Identity: 4. Similarity error: 0.380822 Identity: 4. Similarity error: 0.394788 Identity: 4. Similarity error: 0.433584 Identity: 1. Similarity error: 0.295633 Identity: 1. Similarity error: 0.302817 Identity: 1. Similarity error: 0.310627 Identity: 1. Similarity error: 0.287672 Identity: 1. Similarity error: 0.306451 Identity: 1. Similarity error: 0.322189 Identity: 1. Similarity error: 0.277154 Identity: 1. Similarity error: 0.283235 Identity: 1. Similarity error: 0.294519 Identity: 1. Similarity error: 0.304006 Identity: 9. Similarity error: 0.346696 Identity: 9. Similarity error: 0.297305 Identity: 9. Similarity error: 0.365462 Identity: 9. Similarity error: 0.301169 Identity: 9. Similarity error: 0.334413 Identity: 9. Similarity error: 0.347344 Identity: 9. Similarity error: 0.323791 Identity: 9. Similarity error: 0.348167 Identity: 9. Similarity error: 0.319981 Identity: 9. Similarity error: 0.355038 Identity: 2. Similarity error: 0.307948 Identity: 2. Similarity error: 0.31272 Identity: 2. Similarity error: 0.293307 Identity: 2. Similarity error: 0.30623 Identity: 2. Similarity error: 0.266553 Identity: 2. Similarity error: 0.319812 Identity: 2. Similarity error: 0.285 Identity: 2. Similarity error: 0.300466 Identity: 2. Similarity error: 0.258334 Identity: 2. Similarity error: 0.274881 Identity: 6. Similarity error: 0.465977 Identity: 6. Similarity error: 0.338479 Identity: 6. Similarity error: 0.407876 Identity: 6. Similarity error: 0.38913 Identity: 6. Similarity error: 0.366433 Identity: 6. Similarity error: 0.361704 Identity: 6. Similarity error: 0.424361 Identity: 6. Similarity error: 0.446707 Identity: 6. Similarity error: 0.414173 Identity: 6. Similarity error: 0.4091850.361412

I can mail the complete list if in case it is needed. I am not posting the similarity errors of all 100 true and false test samples as it will not be clear in viewing them in the forum.

I am working on Face recognition as a part of my project. I am using the fisherfaces program provided in OpenCV website for the same. But the program code provided there does not have verification and the recognition always results in a subject in the database even though it is a false subject. The fisherfaces reconstruction does not perform well. The Similarity error for both the true test inputs (true test input - picked from the database) and false test inputs (false test input - NOT picked from the database) are quite close to each other and some even overlap with the range of values from true test inputs.

I performed an experiment on AT & T database which has 40 subjects.

I selected 20 out of them for database and rest 20 i used to provide false test input where the ideal behavior of false test input is not to get recognized to one of the images in database. But the normalized euclidean distance of the false test inputs are quite near and some even overlap with the normalized euclidean distance of the true test inputs. Please let me know if there is any method to set an appropriate threshold and hence make the algorithm work well.

Results for false_test_inputs

Identity: 0. Similarity error: 0.357048 Identity: 0. Similarity error: 0.364141 Identity: 0. Similarity error: 0.368228 Identity: 0. Similarity error: 0.322856 Identity: 0. Similarity error: 0.298744 Identity: 0. Similarity error: 0.3978 Identity: 0. Similarity error: 0.354158 Identity: 8. Similarity error: 0.417423 Identity: 0. Similarity error: 0.42364

Results for true_test_inputs

Identity: 8. Similarity error: 0.291803 Identity: 8. Similarity error: 0.381701 Identity: 8. Similarity error: 0.455005 Identity: 8. Similarity error: 0.378799 Identity: 8. Similarity error: 0.480244 Identity: 8. Similarity error: 0.456196 Identity: 8. Similarity error: 0.419657 Identity: 8. Similarity error: 0.361412

I can mail the complete list if in case it is needed. I am not posting the similarity errors of all 100 true and false test samples as it will not be clear in viewing them in the forum.

I am working on Face recognition as a part of my project. I am using the fisherfaces program provided in OpenCV website for the same. But the program code provided there does not have verification and the recognition always results in a subject in the database even though it is a false subject. The fisherfaces reconstruction does not perform well. The Similarity error for both the true test inputs (true test input - picked from the database) and false test inputs (false test input - NOT picked from the database) are quite close to each other and some even overlap with the range of values from true test inputs.

I performed an experiment on AT & T database which has 40 subjects.

I selected 20 out of them for database and rest 20 i used to provide false test input where the ideal behavior of false test input is not to get recognized to one of the images in database. But the normalized euclidean distance of the false test inputs are quite near and some even overlap with the normalized euclidean distance of the true test inputs. Please let me know if there is any method to set an appropriate threshold and hence make the algorithm work well.

Results for false_test_inputs

Identity: 0. Similarity error: 0.357048 Identity: 0. Similarity error: 0.364141 Identity: 0. Similarity error: 0.368228 Identity: 0. Similarity error: 0.322856 Identity: 0. Similarity error: 0.298744 Identity: 0. Similarity error: 0.3978 Identity: 0. Similarity error: 0.354158 Identity: 8. Similarity error: 0.417423 Identity: 0. Similarity error: 0.42364

Results for true_test_inputs

Identity: 8. Similarity error: 0.291803 Identity: 8. Similarity error: 0.381701 Identity: 8. Similarity error: 0.455005 Identity: 8. Similarity error: 0.378799 Identity: 8. Similarity error: 0.480244 Identity: 8. Similarity error: 0.456196 Identity: 8. Similarity error: 0.419657 Identity: 8. Similarity error: 0.361412

I can mail the complete list if in case it is needed. I am not posting the similarity errors of all 100 true and false test samples as it will not be clear in viewing them in the forum.

I can attach a plot of the samples.At_T.png