Feature extraction using discrete cosine transform and discrimination power analysis with a face recognition technology

2010 ◽  
Vol 43 (4) ◽  
pp. 1431-1440 ◽  
Author(s):  
Saeed Dabbaghchian ◽  
Masoumeh P. Ghaemmaghami ◽  
Ali Aghagolzadeh
Author(s):  
Mourad Moussa ◽  
Maha Hmila ◽  
Ali Douik

Face recognition is a computer vision application based on biometric information for automatic person identification or verification from image sequence or a video frame. In this context DCT is the easy technique to determine significant parameters. Until now the main object is selection of the coefficients to obtain the best recognition. Many techniques rely on premasking windows to discard the high and low coefficients to enhance performance. However, the problem resides in the shape and size of premask. To improve discriminator ability in discrete cosine transform domain, we used fractional coefficients of the transformed images with discrete cosine transform to limit the coefficients area for a better performance system. Then from the selected bands, we use the discrimination power analysis to search for the coefficients having the highest power to discriminate different classes from each other. Feature selection algorithm is a key issue in all pattern recognition system, in fact this algorithm is utilized to define features vector among several ones, where these features are selected according a specified discrimination criterion. Many classifiers are used to evaluate our approach like, support vector machine and random forests. The proposed approach is validated with Yale and ORL Face databases. Experimental results prove the sufficiency of this method in face and facial expression recognition field.


Author(s):  
David Zhang ◽  
Xiao-Yuan Jing ◽  
Jian Yang

This chapter provides a feature extraction approach that combines the discrete cosine transform (DCT) with LDA. The DCT-based frequency-domain analysis technique is introduced first. Then, we describe the presented discriminant DCT approach and analyze its theoretical properties. Finally, we offer detailed experimental results and a chapter summary.


2021 ◽  
Vol 10 (5) ◽  
pp. 2796-2803
Author(s):  
Linggo Sumarno ◽  
Rifai Chai

The conducted research proposes a feature extraction and classification combination method that is used in a tone recognition system for musical instruments. It is expected that by implementing this combination, the tone recognition system will require fewer feature extraction coefficients than those previously investigated. The proposed combination comprises of feature extraction using discrete cosine transform (DCT) and classification using support vector machine (SVM). Bellyra, clarinet, and pianica tones were used in the experiment, with each indicating a tone with one, several, or many major local peaks in the transform domain. Based on the results of the tests, the proposed combination is efficient enough to be used in a tone recognition system for musical instruments. This is indicated in recognizing a tone, it only needs at least eight feature extraction coefficients.


2021 ◽  
Vol 3 (1) ◽  
pp. 1-6
Author(s):  
Zulfrianto Yusrin Lamasigi

Batik merupakan kain yang dibuat khusus, batik sendiri terbilang unik karena memiliki motif tertentu yang dibuat berdasarkan unsur budaya dari daerah asal batik itu dibuat. setiap motif dan warna batik berbeda-beda sehingga sulit untuk dikenali asal dari motir batik itu sendiri. penelitian ini bertujuan untuk meningkatkan hasil ektraksi fitur pada identifikasi motif batik. metode yang digunakan dalam penelitian ini adalah Discrete Cosine Transform bertujuan untuk meningkatkan hasil ektraksi fitur Gray Level Co-Occurrence Matrix untuk mendapatkan hasil akurasi identifikasi motif batik yang lebih baik, sedangkan untuk mengetahui nilai kedekatan antara data training dengan data testing citra batik akan menggunakan K-Nearest Neighbour berdasarkan nilai ekstraksi fitur yang diperoleh. dalam eksperimen ini dilakukan 4 kali percobaan berdasarkan sudut 0°, 45°, 90°, dan 135° pada nilai k=1, 3, 5, 7, dan 9. sementara itu, untuk menghitung tingkat akurasi dari klasifikasi KNN akan menggunakan confusion matrix. Dari uji coba yang di lakukan dengan menggunakan jumalah data training sebanyak 602 citra dan data testing 344 citra terhadap semua kelas berdasarkan sudut 0°, 45°, 90°, dan 135° pada nilai k=1, 3, 5, , dan 9 akurasi tertinggi yang diperoleh DCT-GLCM ada pada sudut 135° dengan nilai k=3 sebesar 84,88% dan yang paling rendah ada pada sudut 0° dengan nilai k=7 dan 9 sebesar 41,86%. Sedangkan hasil uji dengan hanya mennggunakan GLCM akurasi tertinggi ada pada sudut 135° dengan nilai k=1 sebesar 77,90% dan yang paling rendah ada pada sudut 90° dengan nilai k=7 sebesar 40,69%. Dari hasil uji coba yang dilakukan menunjukkan bahwah DCT bekerja dengan baik untuk meningkatkan hasil ekstraksi fitur GLCM yang dibuktikan dengan hasil rata-rata akurasi yang diperoleh.Batik is a specially made cloth, batik itself is unique because it has certain motifs that are made based on cultural elements from the area where the batik was made. each batik motif and color is different so it is difficult to identify the origin of the batik motir itself. This study aims to improve the feature extraction results in the identification of batik motifs. The method used in this research is Discrete Cosine Transform, which aims to increase the extraction of the Gray Level Co-Occurrence Matrix feature to obtain better accuracy results for identification of batik motifs, while to determine the closeness value between training data and batik image testing data will use K- Nearest Neighbor based on the feature extraction value obtained. In this experiment, 4 experiments were carried out based on angles of 0 °, 45 °, 90 °, and 135 ° at values of k = 1, 3, 5, 7, and 9. Meanwhile, to calculate the level of accuracy of the KNN classification, confusion matrix will be used. . From the trials carried out using the total training data of 602 images and testing data of 344 images for all classes based on angles of 0 °, 45 °, 90 °, and 135 ° at values of k = 1, 3, 5, and 9 accuracy The highest obtained by DCT-GLCM was at an angle of 135 ° with a value of k = 3 of 84.88% and the lowest was at an angle of 0 ° with values of k = 7 and 9 of 41.86%. While the test results using only GLCM, the highest accuracy is at an angle of 135 ° with a value of k = 1 of 77.90% and the lowest is at an angle of 90 ° with a value of k = 7 of 40.69%. From the results of the trials conducted, it shows that the DCT works well to improve the results of the GLCM feature extraction as evidenced by the average accuracy results obtained.


Author(s):  
HEYDI MENDEZ-VÁZQUEZ ◽  
JOSEF KITTLER ◽  
CHI HO CHAN ◽  
EDEL GARCÍA-REYES

Variations in illumination is one of major limiting factors of face recognition system performance. The effect of changes in the incident light on face images is analyzed, as well as its influence on the low frequency components of the image. Starting from this analysis, a new photometric normalization method for illumination invariant face recognition is presented. Low-frequency Discrete Cosine Transform coefficients in the logarithmic domain are used in a local way to reconstruct a slowly varying component of the face image which is caused by illumination. After smoothing, this component is subtracted from the original logarithmic image to compensate for illumination variations. Compared to other preprocessing algorithms, our method achieved a very good performance with a total error rate very similar to that produced by the best performing state-of-the-art algorithm. An in-depth analysis of the two preprocessing methods revealed notable differences in their behavior, which is exploited in a multiple classifier fusion framework to achieve further performance improvement. The superiority of the proposal is demonstrated in both face verification and identification experiments.


Sign in / Sign up

Export Citation Format

Share Document