Identifikasi Barcode pada Gambar yang Ditangkap Kamera Digital Menggunakan Metode JST

AbstrakDewasa ini hampir setiap produk konsumen memiliki label barcode. Namun alat pembaca barcode jenis laser memiliki kelemahan karena tidak dapat mengenali barcode yang mengalami goresan atau noise. Namun telah dikembangkan teknik lain dengan memanfaatkan kamera digital untuk identifikasi barcode. JST telah banyak digunakan untuk identifikasi berbagai macam pola. Proses identifikasi barcode dalam JST terdiri dari proses training dan proses identifikasi. Proses training menggunakan metode LVQ (Learning Vector Quantization). Proses identifikasi terdiri dari beberapa tahap, yaitu akuisisi citra, preprocessing, locating barcode, proses pengujian dan verifikasi. Berdasarkan hasil pengujian metode LVQ dapat digunakan untuk identifikasi foto barcode dengan kinerja yang baik. Hasil pengujian menunjukkan tingkat akurasi sebesar 73,6 % dari 72 citra yang diuji dengan waktu rata-rata adalah 0.5 detik. Sementara waktu yang dibutuhkan untuk menemukan lokasi barcode adalah sekitar 6 detik menggunakan blok dengan ukuran 32x32 pixel. Kata kunci— Barcode, Learning Vector Quantization, Jaringan Syaraf Tiruan AbstrakIn today’s modern society, almost every consumer product has a barcode label. But the barcode reader with laser type has the disadvantage of not being able to recognize the barcode has a scratch or noise. However, other techniques have been developed by using a digital camera for barcode identification. ANN has been widely used for identification of various patterns. Barcode identification process consists of the ANN training process and the identification process. Training process using the LVQ (Learning Vector Quantization). Identification process consists of several stages: image acquisition, preprocessing, locating barcode, testing and verification process. Based on test results LVQ method can be used for photo identification barcode with good performance. The test results showed an accuracy of 73.6% rate of 72 images were tested with an average time is 0.5 seconds. While the time required to find the location of the barcode is about 6 seconds using a block size of 32x32 pixels. Keyword— Barcode, Learning Vector Quantization, Artificial Neural Network

Download Full-text

Klasifikasi Jenis Golongan Darah Menggunakan Fuzzy C-Means Clustering (FCM) dan Learning Vector Quantization (LVQ)

MATICS ◽

10.18860/mat.v10i1.5356 ◽

2018 ◽

Vol 10 (1) ◽

pp. 26

Author(s):

Fajar Rohman Hariri

Keyword(s):

Vector Quantization ◽

Learning Vector Quantization ◽

The Body ◽

Blood Type ◽

Test Results ◽

Font Size ◽

Optimal Result ◽

Fuzzy C Means ◽

Fuzzy C Means Clustering ◽

Index Terms

Abstract— Blood is an important part of the body. Blood is divided into several groups A, B, O, and AB. Conventionally, detect blood group by dripping anti-A serum and anti-B serum into the blood to be recognized and direct measurement of the serum droplet reaction. This study will compare the processes that use segmentation and without using segmentation to know the various segmentation information in introduction of human blood type image. From the test results that segmentation increase accuracy of recognition between 10% -24% of each test. By using JST Learning Vector Quantization (LVQ) as a classifier and Fuzzy C-Mean as segmentation, the optimal result on the system averages 92% to 98%.. Index Terms—Blood, Segmentation, Classification Abstrak–- Darah merupakan salah satu bagian penting dalam tubuh. Darah dibedakan menjadi beberapa golongan yaitu A, B, O, dan AB. Secara konvensional, mendeteksi golongan darah dengan cara meneteskan serum anti-A dan serum anti-B ke darah yang akan dikenali kemudian melakukan pengamatan langsung terhadap reaksi tetesan serum tersebut. Penelitian ini akan membandingkan antara proses pengenalan yang menggunakan segmentasi dengan proses pengenalan tanpa menggunakan segmentasi untuk mengetahui seberapa besar pengaruh metode segmentasi dalam pengenalan citra golongan darah manusia. Dari hasil pengujian didapatkan bahwa dengan adanya metode segmentasi akurasi system pengenalan bertambah antara 10%-24% setiap uji coba. Dengan menggunakan JST Learning Vector Quantization (LVQ) sebagai pengklasifikasi dan Fuzzy C-Mean sebagai segmentasi citra darah dapat diperoleh hasil yang optimal pada sistem pengenala golongan darah manusia dengan prosentase keberhasilan rata rata 92% hingga 98%. <a name="PointTmp">Kata Kunci—Darah, Segmentasi, Klasifikasi </a><div><table width="637" cellspacing="0" cellpadding="0"><tbody><tr><td style="padding: 9.35pt;" align="left" valign="top" height="181"><p class="Authors" style="margin-bottom: .0001pt; mso-element: frame; mso-element-frame-width: 468.75pt; mso-element-frame-height: 117.05pt; mso-element-wrap: no-wrap-beside; mso-element-anchor-horizontal: page; mso-element-left: 85.2pt; mso-element-top: 43.85pt; mso-height-rule: exactly;">Klasifikasi Jenis Golongan Darah MenggunakanFuzzy C-Means Clustering (FCM) dan Learning Vector Quantization (LVQ)</td></tr></tbody></table></div>

Download Full-text

Verifikasi Tanda Tangan Menggunakan Ekstraksi Fitur LBP dan Klasifikasi LVQ

Jurnal Teknologi Informasi, Komputer, dan Aplikasinya (JTIKA ) ◽

10.29303/jtika.v2i2.107 ◽

2020 ◽

Vol 2 (2) ◽

pp. 208-216

Author(s):

Medeline Widia Andani ◽

Fitri Bimantoro

Keyword(s):

Computer System ◽

Vector Quantization ◽

Local Binary Pattern ◽

Learning Vector Quantization ◽

Signature Verification ◽

Identification Process ◽

The Media ◽

Verification Process ◽

General Signature

Signature is one of the media used for verification and legalization of information, such as documents that are closely related to legality. In general, signature verification is done manually by direct comparing, this is certainly not effective, especially if doing a lot verification. Therefore, we need a computer system that can automatically verify a person's signature to save time in matching and reducing errors. This research was conducted using feature of Local Binary Pattern (LBP) method and Learning Vector Quantization (LVQ) classifier. Materials that used in this research are 600 signature images with a size of 500x500 pixels taken from 30 respondents where each respondent taken 15 original signatures and 5 fake signatures. The results of this research are that the signature identification process resulted in 93% and the verification process resulted in an accuracy of 63%, a sensitivity of 89%, and a specificity of 42%.

Download Full-text

Identification of Formaldehyde Bananas using Learning Vector Quantization

Indonesian Journal of Information Systems ◽

10.24002/ijis.v3i2.4110 ◽

2021 ◽

Vol 3 (2) ◽

pp. 160

Author(s):

Rahmat Musa ◽

Mutaqin Akbar

Keyword(s):

Vector Quantization ◽

Visual Recognition ◽

Texture Features ◽

Learning Vector Quantization ◽

Learning Rate ◽

Training Data ◽

Blue Color ◽

Test Results ◽

Accurate Identification ◽

Sample Data

Bananas that ripen with chemical process or do not ripen naturally usually, this can be recognized by the presence of blackish patches on the surface of the skin. But visual recognition has its drawbacks, which is that it is difficult to recognize similarities between formalin bananas and natural bananas, resulting in a lack of accurate identification. In this study, a system was built that can determined formalin bananas and natural bananas through digital image identification using supervised classification. The image to be identification previously goes through the process of transforming RGB (Red Green Blue) color to Grayscale, and the process of extracting texture features using statically recognizable features through histograms, in the form of average, standard deviation, skewness, kurtosis, energy, entropy and smoothness. The extraction of texture features is classified with LVQ (Learning Vector Quantization) to determine formalin or natural bananas. The test was conducted with 122 banana imagery sample data, 100 imagery as training data consisting of 50 imagery for natural bananas and 50 imagery for bananas formalin, 22 imagery as test data. The test results showed LVQ method has the best percentage at Learning Rate 0.1, Decreased Learning Rate 0.75 and maximum epoch of 1000 with the smallest epoch of 7, obtained accuracy 90.90%, precision 84.61% and recall 100%.

Download Full-text