Garuda - Garba Rujukan Digital

FACIAL EXPRESSIONS RECOGNITION USING BACKPROPAGATION NEURAL NETWORK FOR MUSIC PLAYLIST ELECTIONS Setiawardhana Setiawardhana; Nana Ramadijanti; Peni Rahayu
Jurnal Ilmiah Kursor Vol 6 No 3 (2012)
Publisher : Universitas Trunojoyo Madura

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

Penelitian ini dibuat untuk mengenali ekspresi wajah sebagai indikator untuk menjalankan playlist musik. Sistem pengenalan ekspresi wajah berasal dari data masukan seseorang yang diambil secara offline, dengan posisi terdekat dengan kamera, dimana posisi wajah tidak boleh miring. Prosesnya dengan pengambilan citra wajah secara offline yang dikenali dengan kombinasi warna, dan mengekstrak fitur penting dari wajah berdasarkan lokasi alis, mata, dan bentuk mulut kemudian mengenali ekspresi wajah menggunakan Jaringan Saraf Tiruan Propagasi Balik (Backpropagation Neural Network). Ekspresi yang akan dikenali Data keluaran dari pengenalan ekspresi wajah berupa indek yang secara otomatis akan digunakan sebagai indikator untuk menjalankan musik, sehingga musik akan berubah mengikuti perubahan ekspresi wajah seseorang. Sistem yang telah dibuat dapat mengenali tiga jenis ekspresi yaitu: normal, marah, dan bahagia. Pengujian dengan pengambilan gambar wajah secara offline sebagai data masukan untuk Jaringan Saraf Tiruan Propagasi Balik, dimana pada saat pembelajaran diperoleh hasil yang konvergen dengan error terendah dengan jumlah neuron pada lapisan hidden sebanyak 10 unit, nilai laju pembelajaran sebesar 0.0625325 dan nilai mean square error sebesar 0.0135. Kata Kunci: Ekspresi Wajah, Backpropagation, Music Playlist. Abstract The objective of the research is to detect facial expression as indicator to cast a music playlist. Facial expression detection system input is performed offline by taking photograph of a subject with nearest position from the camera and facial position should not be tilted. The image is identified as a combination of color and feature extraction is performed based on location of eyebrow, eye, and mouth. Facial expression is detected with Artificial Neural Network Backpropagation method. The output data is an index, which automatically select and play the music. In this way, the music is modified according to the changes of facial expression. The system is designed to detect three facial expressions: normal, angry, and happy expression. The similarity between features values from each expression influence the ability to differentiate each expression. Offline system evaluation is performed with backpropagation neural network method,for learning process, it reaches convergent value with lowest error value when using 10 unit neuron on hidden layer, learning rate value is 0.0625325 and mean square error value is 0.0135.

Optimization of Fuzzy Social Force Model Adaptive Parameter using Genetic Algorithm for Mobile Robot Navigation Control Alif Wicaksana Ramadhan; Bima Sena Bayu Dewantara; Setiawardhana Setiawardhana
Jurnal Rekayasa Elektrika Vol 19, No 1 (2023)
Publisher : Universitas Syiah Kuala

The Social Force Model (SFM) is a popular navigation technique for mobile robots that is primarily used to simulate pedestrian movement. The SFM method's drawback is that several parameter values, such as gain, k, and impact range, σ, must be determined manually. The reaction of the SFM is frequently inappropriate for certain environmental circumstances as a result of this manual determination. In this paper, we propose employing the Fuzzy Inference System (FIS), whose rules are optimized using a Genetic Algorithm (GA) to manage the value of the gain, k, parameter adaptive. The relative distance, d, and relative angle, α, concerning the robot's obstacle are the inputs for the FIS. The test results using a 3-D realistic CoppeliaSim demonstrated that the learning outcomes of FIS rules could provide adaptive parameter values suitable for each environmental circumstance, allowing the robot to travel smoothly is represented using the robot’s heading deviation which decreasing by and reaching the goal 1.6 sec faster from the starting point to the goal, compared to the SFM with the fixed parameter value. So that the proposed method is more effective and promising when deploying on the real robot implementation.

Handling Missing Value dengan Pendekatan Regresi pada Dataset Akuakultur Berukuran Kecil Ricky Afiful Maula; Agus Indra Gunawan; Bima Sena Bayu Dewantara; M. Udin Harun Al Rasyid; Setiawardhana Setiawardhana; Ferry Astika Saputra; Junaedi Ispianto
Jurnal Rekayasa Elektrika Vol 18, No 3 (2022)
Publisher : Universitas Syiah Kuala

Shrimp cultivation is strongly influenced by pond water quality conditions. Farmers must know the appropriate action in regulating water quality that is suitable for shrimp survival. The state of water quality can be understood by measuring pond parameters using various sensors. Installing sensors equipped with artificial intelligence modules to inform water quality conditions is the right action. However, the sensor cannot be separated from errors, so it results in not being able to get data or missing data. In this case, the approach of 5 parameters of pond water quality from 13 available parameters is carried out. This paper proposes a technique to obtain lost data caused by sensor error and looks for the best model. A simple approach can be taken, such as the Handling Missing Value (HMV), which is commonly used, namely the mean, with the K-Nearest Neighbors (KNN) classifier optimized using a grid search. However, the accuracy of this technique is still low, reaching 0.739 at 20-fold cross-validation. Calculations were carried out with other methods to further improve the prediction accuracy. It was found that Linear Regression (LR) can increase accuracy up to 0.757, which outperforms different approaches such as the statistical approach to mean 0.739, mode 0.716, median 0.734, and regression approach KNN 0.742, Lasso 0.751, Passive Aggressive Regressor (PAR) 0.737, Support Vector Regression (SVR) 0.739, Kernel Ridge (KR) 0.731, and Stochastic Gradient Descent (SGD) 0.734.

Optimization of Fuzzy Social Force Model Adaptive Parameter using Genetic Algorithm for Mobile Robot Navigation Control Alif Wicaksana Ramadhan; Bima Sena Bayu Dewantara; Setiawardhana Setiawardhana
Jurnal Rekayasa Elektrika Vol 19, No 1 (2023)
Publisher : Universitas Syiah Kuala

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.17529/jre.v19i1.28330

The Social Force Model (SFM) is a popular navigation technique for mobile robots that is primarily used to simulate pedestrian movement. The SFM method's drawback is that several parameter values, such as gain, k, and impact range, σ, must be determined manually. The reaction of the SFM is frequently inappropriate for certain environmental circumstances as a result of this manual determination. In this paper, we propose employing the Fuzzy Inference System (FIS), whose rules are optimized using a Genetic Algorithm (GA) to manage the value of the gain, k, parameter adaptive. The relative distance, d, and relative angle, α, concerning the robot's obstacle are the inputs for the FIS. The test results using a 3-D realistic CoppeliaSim demonstrated that the learning outcomes of FIS rules could provide adaptive parameter values suitable for each environmental circumstance, allowing the robot to travel smoothly is represented using the robot’s heading deviation which decreasing by and reaching the goal 1.6 sec faster from the starting point to the goal, compared to the SFM with the fixed parameter value. So that the proposed method is more effective and promising when deploying on the real robot implementation.

Handling Missing Value dengan Pendekatan Regresi pada Dataset Akuakultur Berukuran Kecil Ricky Afiful Maula; Agus Indra Gunawan; Bima Sena Bayu Dewantara; M. Udin Harun Al Rasyid; Setiawardhana Setiawardhana; Ferry Astika Saputra; Junaedi Ispianto
Jurnal Rekayasa Elektrika Vol 18, No 3 (2022)
Publisher : Universitas Syiah Kuala

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.17529/jre.v18i3.25903

Shrimp cultivation is strongly influenced by pond water quality conditions. Farmers must know the appropriate action in regulating water quality that is suitable for shrimp survival. The state of water quality can be understood by measuring pond parameters using various sensors. Installing sensors equipped with artificial intelligence modules to inform water quality conditions is the right action. However, the sensor cannot be separated from errors, so it results in not being able to get data or missing data. In this case, the approach of 5 parameters of pond water quality from 13 available parameters is carried out. This paper proposes a technique to obtain lost data caused by sensor error and looks for the best model. A simple approach can be taken, such as the Handling Missing Value (HMV), which is commonly used, namely the mean, with the K-Nearest Neighbors (KNN) classifier optimized using a grid search. However, the accuracy of this technique is still low, reaching 0.739 at 20-fold cross-validation. Calculations were carried out with other methods to further improve the prediction accuracy. It was found that Linear Regression (LR) can increase accuracy up to 0.757, which outperforms different approaches such as the statistical approach to mean 0.739, mode 0.716, median 0.734, and regression approach KNN 0.742, Lasso 0.751, Passive Aggressive Regressor (PAR) 0.737, Support Vector Regression (SVR) 0.739, Kernel Ridge (KR) 0.731, and Stochastic Gradient Descent (SGD) 0.734.

Title

Found 5 Documents
Search

Abstract

Abstract

Abstract

Abstract

Abstract

Title Search

Found 5 Documents Search

Abstract

Abstract

Abstract

Abstract

Abstract

Title

Found 5 Documents
Search