Claim Missing Document
Check
Articles

Metode Ekstraksi Ciri untuk Membedakan Citra Wajah Asli dan Foto Berbasis Perceptron Afri Yudamson , Indah Soesanti, Warsun Najib
Semesta Teknika Vol 16, No 1 (2013): MEI 2013
Publisher : Universitas Muhammadiyah Yogyakarta

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.18196/st.v16i1.431

Abstract

Face is one of media for human identification. Previous studies aimed at identifying human face were for a two-dimensional images. Thus, fraud may occur when providing input in two-dimensional face images (photos). This study aims to distinguish the original three-dimensional face image with two-dimensional face image. Feature extraction based on facial geometry principles (Incomplete sentence, subject only, do not know what the authors mean). Face images (both the original and the photos) were captured at deviated angle, to the left and to the right. Each image is then sliced for each face components (eyes and nose) and sought the position of the center point of each component. Comparison between the value of the right eye-nose projection vector to the left-right eye vector and the value of the left-right eye vector become the characteristics of each image. The perceptron method was used for the classifiers. The result, the software can distinguish the original three-dimensional and two-dimensional face image with an error of 8.33% of the 24 tested images. Error occurred for some samples that show big round nose.
Analisis Kualitas dan Kuantitas Steganografi dengan Interpolasi pada Citra Medis Meirista Wulandari; Indah Soesanti
Forum Teknik Vol 36, No 1 (2015)
Publisher : Faculty of Engineering, Universitas Gadjah Mada

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (9630.534 KB)

Abstract

Nowadays sharing of medical images through Internet is very popular to make telediangosis, telesurgeries and teleconsultation. Steganography is an art of hiding information inside of a media, which is called cover media. This research presents steganography image on medical image. The medical image is divided into two areas edge and non-edge area. Each area has its embedding and extraction algorithm. A medical data in a text form is hidden insidemedical image by interpolation technique. Interpolation technique is used to increase the embedding capacity. After steganography process, medical data and its cover image and its earlier image must be recovered and reversibled from extraction its stego image. The results show an average of the optimum embedding capacity is 103,404 bit with average PSNR is 41.8682 dB. Furthermore, analysis on texture of the stego image is done by this research. The analysis shows the impact of embedding process. The results of texture analysis is embedding process gives a big impact to energy 27.9199%, entropy 5.4725% and skewness 1.5266%.Keywords: steganography, data, medical image, interpolation, texture image
Pengenalan Wajah Dengan Metode Eigenface Urifan, Isbadi; Hidayat, Risanuri; Soesanti, Indah
Proceedings of KNASTIK 2010
Publisher : Duta Wacana Christian University

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Wajah merupakan bagian dari tubuh manusia yang menjadi fokus perhatian di dalam interaksi sosial, Wajah memainkan peranan vital dengan menunjukan identitas seseorang. Oleh karena itu wajah digunakan sebagai organ dari tubuh manusia yang dijadikan indikasi pengenalan seseorang atau face recognition. Salah satu metode pengenalan wajah adalah metode eigenface. Eigenface adalah suatu metode pengenalan wajah yang berdasarkan pada algoritma Principal Component Analysis (PCA). Secara singkat prosesnya adalah citra direpresentasikan dalam sebuah gabungan vektor yang dijadikan satu matriks tunggal. Dari matriks tunggal ini akan diekstraksi suatu ciri utama yang akan membedakan antara citra wajah satu dengan citra wajah lainnya. Citra yang digunakan adalah citra digital dengan format grayscale untuk mempermudah komputasinya. Dengan membandingkan antara citra uji dengan citra referensi menggunakan konsep jarak euclidean, maka akan didapat kesimpulan apakah suatu citra wajah dikenali atau tidak dikenali.
Deteksi Iris Berdasarkan Metode Black Hole dan Circle Curve Fitting Kurnianto, Danny; Soesanti, Indah; Nugroho, Hanung Adi
JURNAL INFOTEL Vol 5 No 2 (2013): November 2013
Publisher : LPPM INSTITUT TEKNOLOGI TELKOM PURWOKERTO

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (574.566 KB) | DOI: 10.20895/infotel.v5i2.3

Abstract

Sistem pengenalan identitas personal berdasarkan ciri biometrika adalah suatu sistem pengenalan seseorang berdasarkan pada ciri biometrika yang melekat pada orang tersebut. Iris mata merupakan salah satu ciri biometrik yang handal untuk sistem pengenalan identitas personal. Bagian sistem pengenalan identitas personal berdasarkan biometrik iris yang dianggap paling krusial adalah deteksi lokasi iris, karena akurasi deteksi iris berpengaruh pada tingkat akurasi sistem secara keseluruhan. Lokasi iris pada citra mata dibatasi oleh dua buah lingkaran yang memisahkan antara bagian iris dengan pupil dan sklera. Telah banyak metodemetode yang diusulkan oleh para peneliti untuk menghasilkan deteksi lokasi iris dengan akurat dan cepat. Masalah akurasi, kecepatan waktu eksekusi dan ketahanan terhadap noise merupakan bidang penelitian yang menantang pada deteksi iris. Makalah ini menyajikan metode deteksi iris menggunakan metode black hole dan circle curve fitting. Langkah pertama, mencari batas dalam lingkaran iris yang memisahkan antara daerah iris dan pupil. Dengan metode black hole yang bekerja berdasarkan fakta bahwa lokasi pupil merupakan daerah lingkaran yang paling hitam dan memiliki distribusi nilai intensitas yang seragam, maka lokasi pupil dapat ditentukan dengan teknik pengambangan. Batas lingkaran pupil dapat ditentukan dengan circle curve fitting dari parameter lingkaran daerah pupil. Langkah kedua, mencari batas luar lingkaran iris yang memisahkan antara iris dan sklera. Peta tepi citra iris dicari dengan menggunakan deteksi tepi Canny, kemudian diambil satu komponen tepi arah vertikal yang dapat mewakili batas lingkaran luar iris. Dari komponen tepi tersebut, dihitung jari-jari iris yang berpusat di pusat pupil. Dengan jari-jari iris dan pusat iris maka dapat ditentukan batas luar iris menggunakan circle curve fitting
EKSTRAKSI CIRI FOVEA AVASCULAR ZONE (FAZ) BERBASIS WAVELET PADA PENDERITA DIABETIC RETINOPATHY Purnamasar, Dewi; Nugroho, Hanung Adi; Soesanti, Indah
Prosiding SNATIF Vol 1, No 1 (2014): Prosiding Seminar Nasional Teknologi dan Informatika
Publisher : Prosiding SNATIF

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Abstrak  Jenis Diabetic Retinopathy (DR) adalah komplikasi okular yang paling umum dan serius dari Diabetes Mellitus (DM) yang mengganggu retina. Komplikasi ini menyebabkan kebutaan. Faktor yang menentukan DR adalah Fovea Avascular Zone (FAZ). Untuk mengetahui karakteristik dari FAZ  dengan kasat mata sangat susah, karena letaknya berada di daerah makula dan tertutup pembuluh darah vessel. Tujuan dari penelitian ini adalah untuk mengetahui ekstraksi ciri FAZ dengan membandingkan wavelet db2,db9,symlet dan coif1 untuk mendapatkan nilai entropy maupun energi serta untuk mengetahui nilai keakuratan dari masing-masing level penderita DR dengan mata normal. Metode penelitian ini menggunakan wavelet, data base yang digunakan adalah citra retina messidor. Dari hasil penelitian yang telah dilakukan dapat diketahui bahwa wavelet coif1 mempunyai akurasi yang lebih tinggi dibandingkan dengan db2,db9 dan wavelet symlet. Wavelet coif1 menunjukkan tingkat error kesalahan bernilai 21,53%, akurasinya 78,46%. Akurasinya lebih tinggi dibandingkan dengan wavelet yang lain. Hal ini menunjukkan bahwa wavelet coif1 dapat membedakan FAZ mata normal dengan penderita DR. Kata kunci: entropy, Fovea Avascular Zone, vessel, wavelet.
Perancangan Perangkat Lunak untuk Ekstraksi Ciri dan Klasifikasi Pola Batik Soesanti, Indah
Semesta Teknika Vol 17, No 2 (2014): NOVEMBER 2014
Publisher : Semesta Teknika

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

The popularity of batik patterns in Indonesia has varied. Industrial modern devices in imaging have supported batik pattern recognition and classification. The important of product pattern information could not naturally visible. The information about batik pattern can be achieved by using the appropriate software design of image processing for extracting the features. One of the potential procedures is the unsupervised classification method based on specific feature.  In this research, the specific feature extraction based on the eigenimage of batik pattern was done. In the final step, the nearest distance eigenimage between reference batik image and test batik image was used to identify the batik from the classical pattern field point of view. The results of batik image identification conformed 96.67% with the reference batik images.
Evaluasi Nilai Noise Sebelum Dan Sesudah Kalibrasi Sebagai Salah Satu Wujud Kinerja Pesawat CT-Scan Andrey Nino Kurniawan Nino Kurniawan; Indah Soesanti Soesanti
Forum Teknik Vol 33, No 3 (2010)
Publisher : Faculty of Engineering, Universitas Gadjah Mada

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

The noise has to be measured daily because it directly degrades image quality, which inturn it is dificult to define between normal and pathological tissues on a-CT image. Thisresearch’s aim was performed to examine noise differences between before and aftercalibrating of the CT-scanner system, and also to study the chance of noise deviated values day-by-day during 20 days of the two different periods of measurement.A callibrated head water phantom was scanned before and after equipment callibrations,using the head scanning parameter, and device measurement at ROI 228.2 mm2 to obtain thenoise data. Descriptive statistics was employed to present related information, and the T-testpaired-sample was the statistical tool to test the null hypothese (Ho) with level of significance(α) 0.05.The results showed most of noise values between before and after equipment calibrationsfall arround the base line or within two standard deviation (2SD). In before callibration of thenoise, the base line noise value is 4.76 HU whereas the upper limit and the lower limmit of thenoise values are 5.34 HU and 4.18 HU (respectively) with a-1.16 HU difference. In aftercallibration of the noise, the base line noise value is 4.35 HU whereas the upper limit and thelower limmit of the noise values are 5.29 HU and 3.41 HU with a-1.87 HU difference. However,there are two out of fourty noise values (± 5%) that fall outside of the 2SD limmit if a numberfalls outside of 2SD from the base line in the same direction (i.e., all high or all low) more thanfour days in a row, it is a shift that could be due to a machine malfunction and should beinvestigated before the outer control limit is exceeded. Statistical analyses with paired- sampleT-test showed p-value (0.005) < 0.05. By this meaning that, the noise between before and afterCT-Scanner callibrations are different in their values additonally most of the noise values aftercallibration seem to be low compared with that of the values before callibrations.Keywords: CT scan, evaluation, noise.
IDENTIFIKASI TINGKAT KESEGARAN DAGING AYAM BROILER BERDASAR CIRI TEKSTUR DAN WARNA DAGING Prima, Widyawati; Oyas, Wahyu; Indah, Soesanti
Al-Mabsut Vol 6, No 1 (2013): (APRIL 2013)
Publisher : Sekolah Tinggi Agama Islam Ngawi

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

IDENTIFIKASI TINGKAT KESEGARAN DAGING AYAM BROILER BERDASAR CIRI TEKSTUR DAN WARNA DAGING Prima Widyawati W (Mahasiswa Pascasarjana Teknik Elektro, Universitas Gadjah Mada Yogyakarta), Oyas Wahyu N (Staf Pengajar Pascasarjana Teknik Elektro, Universitas Gadjah Mada Yogyakarta), Indah Soesanti (Staf Pengajar Pascasarjana Teknik Elektro, Universitas Gadjah Mada Yogyakarta). Abstrak Tujuan utama dari penelitian ini adalah mengidentifikasi tingkat kesegaran daging ayam broiler berdasar ciri warna dan tekstur. Sistem ini juga dapat digunakan untuk membedakan daging ayam hasil sembelihan yang sesuai syariat Islam dengan daging ayam bangkai atau tiren. Metode identifdikasi tingkat kesegaran daging ayam yang digunakan pada penelitian ini adalah menggunakan pengolahan citra digital yaitu ekstrasi ciri warna metode histogram dan ekstrasi ciri tekstur metode box counting. Sampel daging yang digunakan diambil dari daging ayam  broiler bagian dada. Bagian dada memiliki komposisi terbesar pada tubuh ayam sehingga pengujian pada bagian dada bisa mewakili semua bagian pada ayam. Berdasar hasil penelitian dapat diketahui bahwa dengan menggunakan ciri tekstur dan warna daging dapat dibedakan antara daging ayam segar, kurang segar, dan daging busuk. Berdasar hasil pengujian dapat diketahui ciri warna dapat memberikan hasil identifikasi yang lebih baik dibandingkan ciri tekstur citra daging ayam tersebut. Kata-kata kunci: ekstrasi ciri, RGB, piksel, histogram., box counting
Klasifikasi Wajah Kambing Peranakan Ettawa (PE) Jantan Berbasis Perseptron Soesanti, Indah; Soesanto, Adhi; Chamim, Anna Nur Nazilah
Semesta Teknika Vol 17, No 2 (2014): NOVEMBER 2014
Publisher : Semesta Teknika

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Goat Peranakan Ettawa ( PE ) is a kind of superior goat derived from goat crosses, between Ettawa (Jamnapari ) from India and Kambing Kacang (Bean Goat) from Java. A factor to determine quality of goat PE is it’s face. More than 30 cm ears length and the head color is black represents good quality. More better the quality of goat face, means higher selling price. In this study, male goat face is classified into class good quality, less good, and not good at data such as photo / image In the market, classification done by visual observation, so many farmers have difficulty in classifying the face of a goat. For that purpose, a system is needed that capable for classifying a goat face to facilitate farmers in classifying.This classification system uses Perceptron Method, is a method of guided learning using characteristic as input those are ears length, black value and brown face value. Images are used as training images as much as 9 images, and test images are 20 images. This system could classificating PE goat face with success rate of 95% and 1 error from 20 testing images. Error occured because the background was detected as black and image taking that not precise.
Identifikasi Titik Api Lilin Berbasis Nilai HSV , Threshold dan Momen Citra untuk Aplikasi Robot Pemadam Api Wiyagi, Rama Okta; Soesanti, Indah; Susanto, Adhi
Jurnal Semesta Teknika Vol 17, No 1 (2014): MEI 2014
Publisher : Jurnal Semesta Teknika

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Fire fighting robot is robot that has function to find and extinguish a candle flame in the space arena. To be able carry out their duties then the robot is equipped with sensors, controllers and drivers. Phototransistor, thermopile arrays, or UVTron is sensors that usually used in fire fighting robot. These sensors have some drawbacks. Phototransistor has a relatively close distance readings. While TPA81 thermopile array has a narrow field of reading only 41 ° x 6 ° from sensor. UVtron only limited to determine whether there is any point of the fire and was unable to determine absolute position or angle of the hotspots and vulnerable to damage if the jar is touched by the hand. Additionally TPA81 sensors and sensor UVtron is relatively expensive. This research aims to build a candle light detection alternative better in terms of specification, performance, price, reliability and ease of development. As the input of the system identification using webcams camera types. The webcam running on Raspberry Pi single-board computer. Image information is converted to HSV color space (Hue, Saturation, Value) and applied threshold processing. Thresholding HSV performed on the range of values contained in the object candle flame. To get the absolute position of a candle flame using moments analysis. Identification system can identify candle flame spot with the farthest distance is 225cm. Angle readings in the horizontal plane by 60 ˚ and the vertical plane by 40 ˚. The achievement of the highest FPS obtained in image resolution size of 320 x 240 pixels which is 8.129 FPS.