Dynamic Time Warping Application for Financial Pattern Recognition

2020 ◽  
Author(s):  
Yangqi Li ◽  
Taihua Hu
Author(s):  
Ersa Triansyah ◽  
Youllia Indrawaty N

[Id] Pattern recognition memiliki kemampuan untuk mengenali suara dengan melakukan pengenalan pola suara melalui fitur-fitur sinyal suara yang kemudian dilakukan pengenalan pola melalui perbandingan pola suara uji dengan suara referensi. Untuk mendapatkan fitur-fitur sinyal suara, diperlukan metode untuk mengekstraksi sinyal suara sehingga fitur-fitur sinyal suara yang dibutuhkan terpenuhi. MFCC (Mel Frequency Cepstral Coefficients) merupakan alternatif metode untuk melakukan ektraksi sinyal yang menghasilkan koefisien cepstral dari sinyal suara. Koefisien cepstral sinyal suara dari hasil ektraksi tersebut, kemudian dilakukan perbandingan kesesuaian antara suara uji dan suara referensi. DTW (Dynamic Time Warping) salah satu algoritma untuk dapat melakukan perbandingan koefisien tersebut. Dalam kasus pegenalan ucapan huruf hijaiyyah umumnya dilakukan secara talaqqi (belajar intensif) antar seorang guru dengan murid, penilaian yang dilakukan bersifat subjektif berdasarkan kemampuan indera dari seorang guru, untuk itu aplikasi pengucapan huruf hijaiyyah merupakan salah satu alternatif untuk mengenali dan menguji kesesuaian ucapan secara objektif melalui penghitungan matematis dengan melakukan pengenalan pola suara. Dari pengujian yang telah dilakukan, dari 6 orang yang diuji melakukan pengucapan 29 huruf 3 tanda baca dan pengulang sebanyak 5 kali menghasilkan persentase kecocokan suara mencapai di atas 90 %, nilai threshold 1,3 Kata kunci: Speech Recognition, Pattern Recognition, MFCC, DTW, Hijaiyyah [En] Pattern recognition has ability to recognize voice by voice pattern recognition through voice signal features which then carried out voice pattern recognition through comparison of tester voice pattern with a reference voice. To get the sound signal features, it needs a method for extracting sound signal so that required sound signals features are fulfilled. MFCC is an alternative method to perform signal extraction which is produce cepstral coefficients of the sound signal. Cepstral coefficients of sound signal from the extraction then will be compared by the match between tester voice and reference voice. DTW is one of algorithm to do a comparison of the coefficients. In the case of introducing hijaiyyah generally talaqqi (intensively) conducted between a teacher and students, the appraisal is subjective based on the sensory capabilities of the teacher, therefore hijaiyyah pronunciation application is an alternative to identify and test the suitability of speech objectively through mathematical calculations by performing voice pattern recognition. From the testing that has been done, from 6 people tested do pronunciations 29 letters and punctuation repeater 3 to 5 times the yield percentage matches the sound reaches above 90%, a threshold value of 1.3.


Author(s):  
K. C. SANTOSH ◽  
BART LAMIROY ◽  
LAURENT WENDLING

In this paper, we present a pattern recognition method that uses dynamic programming for the alignment of Radon features. The key characteristic of the method is to use dynamic time warping (DTW) to match corresponding pairs of the Radon features for all possible projections. Thanks to DTW, we avoid compressing the feature matrix into a single vector which would otherwise miss information. To reduce the possible number of matchings, we rely on a initial normalization based on the pattern orientation. A comprehensive study is made using major state-of-the-art shape descriptors over several public datasets of shapes such as graphical symbols (both printed and hand-drawn), handwritten characters and footwear prints. In all tests, the method proves its generic behavior by providing better recognition performance. Overall, we validate that our method is robust to deformed shape due to distortion, degradation and occlusion.


Sign in / Sign up

Export Citation Format

Share Document