yuli sun hariyani | Telkom University (original) (raw)

Papers by yuli sun hariyani

Research paper thumbnail of End-to-end Convolutional Neural Network Design for Automatic Detection of Influenza Virus

IEIE Transactions on Smart Processing and Computing, 2021

Research paper thumbnail of Deep Learning-Based Optimal Smart Shoes Sensor Selection for Energy Expenditure and Heart Rate Estimation

Sensors, 2021

Wearable technologies are known to improve our quality of life. Among the various wearable device... more Wearable technologies are known to improve our quality of life. Among the various wearable devices, shoes are non-intrusive, lightweight, and can be used for outdoor activities. In this study, we estimated the energy consumption and heart rate in an environment (i.e., running on a treadmill) using smart shoes equipped with triaxial acceleration, triaxial gyroscope, and four-point pressure sensors. The proposed model uses the latest deep learning architecture which does not require any separate preprocessing. Moreover, it is possible to select the optimal sensor using a channel-wise attention mechanism to weigh the sensors depending on their contributions to the estimation of energy expenditure (EE) and heart rate (HR). The performance of the proposed model was evaluated using the root mean squared error (RMSE), mean absolute error (MAE), and coefficient of determination (R2). Moreover, the RMSE was 1.05 ± 0.15, MAE 0.83 ± 0.12 and R2 0.922 ± 0.005 in EE estimation. On the other hand...

Research paper thumbnail of Perbandingan Suara Paru Normal Dan Abnormal Menggunakan Probabilistic Neural Network Dan Support Vector Machine

eProceedings of Engineering, Apr 1, 2017

Paru-paru adalah organ sistem pernapasan yang berfungsi untuk menukar oksigen dari udara dengan k... more Paru-paru adalah organ sistem pernapasan yang berfungsi untuk menukar oksigen dari udara dengan karbondioksida dari darah dengan bantuan hemoglobin. Sebagai organ yang penting perannya untuk tubuh tentu harus dijaga kesehatannya dari penyakit. Salah satu cara dokter mendiagnosa penyakit pada paru-paru adalah dengan mendengarkan suara pernapasan pada paru-paru dengan stetoskop. Suara paru-paru yang dihasilkan pada beberapa kasus penyakit berbeda-beda sehingga terdapat pola tertentu yang bisa dikenali. Pola suara ini dapat digunakan untuk mengklasifikasikan jenis-jenis penyakit pada paru-paru. Masalah yang timbul adalah suara paru-paru menempati frekuensi yang cukup rendah sekitar 20-2000 Hz, amplitudo yang rendah, masalah kebisingan lingkungan, kepekaan telinga dan pola suara yang mirip antara jenis suara paru-paru yang satu dengan yang lain. Karena faktor-faktor tersebut diatas, kesalahan diagnosis biasa terjadi apabila prosedur auskultasi tidak dilakukan dengan benar. Pada penelitian ini dilakukan perbandingan klasifikasi suara paru-paru normal dan suara paru-paru abnormal untuk menentukan metode terbaik diantara dua metode yang akan digunakan. Suara paru-paru akan didekomposisi dengan menggunakan metode wavelet Daubechies2 level 5. Pada proses klasifikasi akan menggunakan metode klasifikasi Support Vector Machine One Against All, Support Vector Machine One Against One, dan klasifikasi jaringan syaraf tiruan PNN (Probabilistic Neural Network). Hasil penelitian ini adalah: akurasi rata-rata Support Vector Machine One against all sebesar 47,55%, maksimum akurasi sebesar 70%, dan rata-rata waktu komputasi selama 0.006 detik. Akurasi rata-rata Support Vector Machine One Against One sebesar 50.92%, maksimum akurasi sebesar 75%, dan rata-rata waktu komputasi selama 0.012 detik. Akurasi rata-rata Probabilistic Neural Network 70%, maksimum akurasi sebesar 70%, dan rata-rata waktu komputasi selama 0.313 detik.

Research paper thumbnail of Indonesian Vehicle Plate Recognition and Identification Based on Digital Image Processing and Artificial Neural Network

every vehicle has its own license which is given legally. In Indonesia, vehicle registration numb... more every vehicle has its own license which is given legally. In Indonesia, vehicle registration number is an important thing used by system such as: parking system, building security system and toll system. This research is doing a vehicle plate recognition and identification which will read the characters on the plate. The input comes from a real time video. There are six main processes in this system. In preprocessing, the system will enhance frame by using top hat and bottom hat transformation. While in detection, it will detect and crop the plate by using integral protection. After getting the plate position, an identification process will do by doing segmentation in each character. Then a feature extraction which combined with Artificial Neural Network is done to get the character of the vehicle registration number. The best Artificial Neural Network accuracy for character identification is obtained by using hexagonal topology and box distance with average 97,81 % accuracy.

Research paper thumbnail of Diaphysis Fracture on Tibia and Fibula Detection Based on Digital Image Processing and Scan Line Algorithm

IFMBE Proceedings, 2014

The X-Ray images of tibia and fibula are the important aiding for clinical diagnosis of fracture ... more The X-Ray images of tibia and fibula are the important aiding for clinical diagnosis of fracture because detection fracture performed by medical practice based on it. Under conditions of tired eyes, some medical practice miss fracture case. There are many previous research developed various methods to process X-Ray images and make system that can detect fracture automatically. In this paper, we proposes the simple system that can detect fracture on the tibia and fibula in two stages. The first stage image pre-processing using image enhancement, edge detection, filtering to remove noise and results the edge of image perfectly. The second stage is feature extraction using scan line algorithm that find maximum value of the difference distance to the nearest pixels from the right and left border margin of image. The maximum value of scan line is used as a threshold to classify normal cases or fracture cases. Accuracy is calculated based on images were tested right images for all images tested. Total images that used are 70 images, 30 normal images and 40 fracture images. Accuracy in this simulation reach 100% for normal images and 90% for fracture images, with time computation about 2.33 seconds.

Research paper thumbnail of Deteksi dan Klasifikasi Kelainan Jantung Berdasarkan Sinyal Elektrokardiograf Menggunakan Wavelet dan Jaringan Syaraf Tiruan Self Organizing Maps

Research paper thumbnail of End-To-End Deep Learning Architecture for Continuous Blood Pressure Estimation Using Attention Mechanism

Sensors

Blood pressure (BP) is a vital sign that provides fundamental health information regarding patien... more Blood pressure (BP) is a vital sign that provides fundamental health information regarding patients. Continuous BP monitoring is important for patients with hypertension. Various studies have proposed cuff-less BP monitoring methods using pulse transit time. We propose an end-to-end deep learning architecture using only raw signals without the process of extracting features to improve the BP estimation performance using the attention mechanism. The proposed model consisted of a convolutional neural network, a bidirectional gated recurrent unit, and an attention mechanism. The model was trained by a calibration-based method, using the data of each subject. The performance of the model was compared to the model that used each combination of the three signals, and the model with the attention mechanism showed better performance than other state-of-the-art methods, including conventional linear regression method using pulse transit time (PTT). A total of 15 subjects were recruited, and ...

Research paper thumbnail of Feature Analysis of Smart Shoe Sensors for Classification of Gait Patterns

Sensors

Gait analysis is commonly used to detect foot disorders and abnormalities such as supination, pro... more Gait analysis is commonly used to detect foot disorders and abnormalities such as supination, pronation, unstable left foot and unstable right foot. Early detection of these abnormalities could help us to correct the walking posture and avoid getting injuries. This paper presents extensive feature analyses on smart shoes sensor data, including pressure sensors, accelerometer and gyroscope signals, to obtain the optimum combination of the sensors for gait classification, which is crucial to implement a power-efficient mobile smart shoes system. In addition, we investigated the optimal length of data segmentation based on the gait cycle parameters, reduction of the feature dimensions and feature selection for the classification of the gait patterns. Benchmark tests among several machine learning algorithms were conducted using random forest, k-nearest neighbor (KNN), logistic regression and support vector machine (SVM) algorithms for the classification task. Our experiments demonstrat...

Research paper thumbnail of Deteksi Penyakit Covid-19 Berdasarkan Citra X-Ray Menggunakan Deep Residual Network

ELKOMIKA: Jurnal Teknik Energi Elektrik, Teknik Telekomunikasi, & Teknik Elektronika

ABSTRAKPenyakit Coronavirus-2019 atau Covid-19 telah menjadi pandemi global dan menjadi masalah u... more ABSTRAKPenyakit Coronavirus-2019 atau Covid-19 telah menjadi pandemi global dan menjadi masalah utama yang harus segera dikendalikan. Salah satu cara yang dapat dilakukan adalah memutus rantai penyebaran virus tersebut dengan melakukan deteksi dan melalukan karantina. Pencitraan X-Ray dapat dijadikan alternatif dalam mempelajari Covid-19. X-Ray dianggap mampu menggambarkan kondisi paru-paru pada pasien Covid-19 dan dapat menjadi alat bantu diagnosa klinis. Pada penelitian ini, kami mengusulkan pendekatan deep learning berbasis residual deep network untuk deteksi Covid-19 melalui citra chest X-Ray. Evaluasi yang dilakukan untuk mengetahui performa metode yang diusulkan berupa precision, recall, F1, dan accuracy. Hasil eksperimen menunjukkan bahwa usulan metode ini memberikan precision, recall, F1 dan accuracy masing-masing 0,98, 0,95, 0,97 dan 99%. Pada masa mendatang, studi ini diharapkan dapat divalidasi dan kemudian digunakan untuk melengkapi diagnosa klinis oleh dokter.Kata kunci...

Research paper thumbnail of Optimization of Deep Neural Networks for Heartrate Estimation from Face Video Stream to Implement Smart Health-City

The Journal of Korean Institute of Communications and Information Sciences

Research paper thumbnail of Multi-wavelet level comparison on compressive sensing for MRI image reconstruction

Bulletin of Electrical Engineering and Informatics

In this study, we proposed compressive sampling for MRI reconstruction based on sparse representa... more In this study, we proposed compressive sampling for MRI reconstruction based on sparse representation using multi-wavelet transformation. Comparing the performance of wavelet decomposition level, which are Level 1, Level 2, Level 3, and Level 4. We used gaussian random process to generate measurement matrix. The algorithm used to reconstruct the image is . The experimental results showed that the use of wavelet multi-level can generate higher compression ratio but requires a longer processing time. MRI reconstruction results based on the parameters of the peak signal to noise ratio (PSNR) and structural similarity index measure (SSIM) show that the higher the level of decomposition in wavelets, the value of both decreases.

Research paper thumbnail of DA-Capnet: Dual Attention Deep Learning Based on U-Net for Nailfold Capillary Segmentation

IEEE Access

Automatic nailfold capillary segmentation is a challenging task owing to noise and large variabil... more Automatic nailfold capillary segmentation is a challenging task owing to noise and large variabilities in images caused by insufficient focusing and low visibility of the capillaries. This task can be useful to detect and estimate the severity of autoimmune diseases of connective tissues or learning the status of white blood cells based on the cells' blood flow on the nailfold capillary. Previous studies have addressed this task using manual, semi-automated, and automated segmentation method. However, further improvement is still required. With the recent progress of deep learning on medical imaging, we herein propose dual attention deep learning based on U-Net for nailfold capillary segmentation, named DA-CapNet. Our DA-CapNet improves the U-Net architecture by integrating a dual attention module that can capture a better representation of feature maps from input images. Furthermore, DA-CapNet is compared with three baselines: adaptive Gaussian algorithm, SegNet, the original U-Net. We experimentally demonstrate that our proposed method outperforms these baselines.

Research paper thumbnail of Perancangan Aplikasi Pembaca Warna Untuk Penderita Buta Warna Berbasis Android

Jurnal Elektro dan Telekomunikasi Terapan

Penyandang cacat buta warna selalu mengalami kesulitan dalam hal membedakan warna dalam kehidupan... more Penyandang cacat buta warna selalu mengalami kesulitan dalam hal membedakan warna dalam kehidupan sehari-harinya, sehingga dibutuhkan alat bantu bagi mereka. Dalam penelitian ini dibuat sebuah algorithma yang dapat berfungsi untuk membedakan warna, algorithma tersebut diaplikasikan kedalam sebuah android. Aplikasi ini membantu dalam hal membedakan warna (merah, hijau dan biru) dengan pola image processing. Output dari aplikasi ini berupa text dan voice yang merupakan hasil deteksi dari input (captured image) secara langsung menggunakan smartphone android. Hasil yang diperoleh dari penelitian ini, yaitu tingkat akurasi total sistem 85% dengan waktu komputasi 6,67 detik dan jarak terjauh 20 cm.

Research paper thumbnail of Perancangan Dan Implementasi Tuner Gitar Otomatis Dengan Penggerak Motor Servo Berbasis Arduino

Jurnal Elektro dan Telekomunikasi Terapan

Gitar merupakan salah satu alat musik yang paling digemari, dan dapat dimainkan oleh semua orang.... more Gitar merupakan salah satu alat musik yang paling digemari, dan dapat dimainkan oleh semua orang. Saat ini masih banyak ditemukan pengguna gitar yang tidak bisa melakukan tuning senar gitar dan jika dilakukan manual juga membutuhkan waktu yang lama untuk mendapatkan hasil tuning yang akurat. Hal ini dapat diatasi dengan perkembangan teknologi Pengguna gitar sekarang ini dapat melakukan tuning senar lebih mudah dan lebih cepat dengan tingkat toleransi kesalahan frekuensi tuning ± 1 Hz dengan menggunakan tuner gitar otomatis.Perancangan alat tuner gitar otomatis ini menggunakan 5 komponen pembentuk alat yaitu selector switch untuk memilih frekuensi yang akan diatur, Op amp untuk menguatkan amplitudo gelombang suara agar diterima arduino dengan baik, Arduino untuk menginisialisasi frekuensi dan memberi sinyal ke motor servo, Motor servo untuk melakukan tuning senar gitar, dan LED sebagai indikator tuning.Hasil dari penelitian ini adalah keakuratan tuning untuk setiap senar didapatkan ...

Research paper thumbnail of Realisasi Pengendali Intensitas Cahaya Lampu Dengan Kontrol Suara Dan Google Android Speech Recognition Api

Jurnal Elektro dan Telekomunikasi Terapan

Dalam kehidupan sehari-hari setiap orang tidak dapat lepas dari cahaya untuk melakukan berbagai a... more Dalam kehidupan sehari-hari setiap orang tidak dapat lepas dari cahaya untuk melakukan berbagai aktivitas di ruangan. Namun cahaya di ruangan harus disesuaikan dengan kebutuhan aktivitas masing-masing untuk menghindari pemborosan energi listrik. Terkadang sebagian orang sering lupa dan malas untuk mematikan atau menyalakan lampu dalam keadaan tertentu karena saklar yang digunakan masih manual. Oleh karena itu dibutuhan saklar yang dapat dikendalikan dari jarak jauh. Penelitian ini merealisasikan sistem pengendali lampu jarak jauh berbasis suara yang didukung oleh google voice recognition engine dan menggunakan android , tidak hanya mematikan dan menghidupkan lampu, namun juga mengatur lampu dengan beberapa level intensitas cahaya. Arduino Uno R3 digunakan sebagai piranti pengendali dengan Bluetooth sebagai media komunikasi antara smartphone dan mikrokontroller. Berdasarkan pengujian sistem dan alat yang direalisasikan, untuk pengujian di dalam ruangan dengan penghalang, maksimal jan...

Research paper thumbnail of Routing Implementation Based-On Software Defined Network Using Ryu Controller and Openvswitch

Jurnal Teknologi, 2016

Open Flow is a standard protocol for differentiating forward function and control functions to fa... more Open Flow is a standard protocol for differentiating forward function and control functions to facilitate the management of big network of SDN. The research have been carried out before using the emulator SDN Mininet. However Mininet has many shortcomings, such as the performance of which is less than the maximum due to simulation. Then some researchers also use the Net-FPGA as device. This device is less suitable for small scale because the prices are quite expensive and programming is quite complicated. In this study, SDN implementation carried out using OpenvSwitch as forwarding function mounted on TP-Link that has modificated using openwrt as firmware and Raspberry Pi with Ryu SDN Controller as control functions. The result shows that routing static can be implemented on SDN Network which use Raspberry Pi with Ryu Controller as control function with average bandwith 536.0909 Mbits/sec and average uptime network is 10.45 second.

Research paper thumbnail of Pengenalan Plat Nomor Kendaraan Indonesia Berbasis Pengolahan Citra Digital Dan Jaringan Syaraf Tiruan Self-Organizing Map (Som)

Research paper thumbnail of Aplikasi Pendeteksi Penyakit Filariasis Berbasis Citra Darah

Research paper thumbnail of End-to-End Deep Learning Architecture for Continuous Blood Pressure Estimation Using Attention Mechanism

Sensors, Apr 20, 2020

Blood pressure (BP) is a vital sign that provides fundamental health information regarding patien... more Blood pressure (BP) is a vital sign that provides fundamental health information regarding patients. Continuous BP monitoring is important for patients with hypertension. Various studies have proposed cuff-less BP monitoring methods using pulse transit time. We propose an end-to-end deep learning architecture using only raw signals without the process of extracting features to improve the BP estimation performance using the attention mechanism. The proposed model consisted of a convolutional neural network, a bidirectional gated recurrent unit, and an attention mechanism. The model was trained by a calibration-based method, using the data of each subject. The performance of the model was compared to the model that used each combination of the three signals, and the model with the attention mechanism showed better performance than other state-of-the-art methods, including conventional linear regression method using pulse transit time (PTT). A total of 15 subjects were recruited, and electrocardiogram, ballistocardiogram, and photoplethysmogram levels were measured. The 95% confidence interval of the reference BP was [86.34, 143.74] and [51.28, 88.74] for systolic BP (SBP) and diastolic BP (DBP), respectively. The R 2 values were 0.52 and 0.49, and the mean-absolute-error values were 4.06 ± 4.04 and 3.33 ± 3.42 for SBP and DBP, respectively. In addition, the results complied with global standards. The results show the applicability of the proposed model as an analytical metric for BP estimation.

Research paper thumbnail of Automated White Blood Cell Counting in Nailfold Capillary Using Deep Learning Segmentation and Video Stabilization

Sensors

White blood cells (WBCs) are essential components of the immune system in the human body. Various... more White blood cells (WBCs) are essential components of the immune system in the human body. Various invasive and noninvasive methods to monitor the condition of the WBCs have been developed. Among them, a noninvasive method exploits an optical characteristic of WBCs in a nailfold capillary image, as they appear as visual gaps. This method is inexpensive and could possibly be implemented on a portable device. However, recent studies on this method use a manual or semimanual image segmentation, which depends on recognizable features and the intervention of experts, hindering its scalability and applicability. We address and solve this problem with proposing an automated method for detecting and counting WBCs that appear as visual gaps on nailfold capillary images. The proposed method consists of an automatic capillary segmentation method using deep learning, video stabilization, and WBC event detection algorithms. Performances of the three segmentation algorithms (manual, conventional, ...

Research paper thumbnail of End-to-end Convolutional Neural Network Design for Automatic Detection of Influenza Virus

IEIE Transactions on Smart Processing and Computing, 2021

Research paper thumbnail of Deep Learning-Based Optimal Smart Shoes Sensor Selection for Energy Expenditure and Heart Rate Estimation

Sensors, 2021

Wearable technologies are known to improve our quality of life. Among the various wearable device... more Wearable technologies are known to improve our quality of life. Among the various wearable devices, shoes are non-intrusive, lightweight, and can be used for outdoor activities. In this study, we estimated the energy consumption and heart rate in an environment (i.e., running on a treadmill) using smart shoes equipped with triaxial acceleration, triaxial gyroscope, and four-point pressure sensors. The proposed model uses the latest deep learning architecture which does not require any separate preprocessing. Moreover, it is possible to select the optimal sensor using a channel-wise attention mechanism to weigh the sensors depending on their contributions to the estimation of energy expenditure (EE) and heart rate (HR). The performance of the proposed model was evaluated using the root mean squared error (RMSE), mean absolute error (MAE), and coefficient of determination (R2). Moreover, the RMSE was 1.05 ± 0.15, MAE 0.83 ± 0.12 and R2 0.922 ± 0.005 in EE estimation. On the other hand...

Research paper thumbnail of Perbandingan Suara Paru Normal Dan Abnormal Menggunakan Probabilistic Neural Network Dan Support Vector Machine

eProceedings of Engineering, Apr 1, 2017

Paru-paru adalah organ sistem pernapasan yang berfungsi untuk menukar oksigen dari udara dengan k... more Paru-paru adalah organ sistem pernapasan yang berfungsi untuk menukar oksigen dari udara dengan karbondioksida dari darah dengan bantuan hemoglobin. Sebagai organ yang penting perannya untuk tubuh tentu harus dijaga kesehatannya dari penyakit. Salah satu cara dokter mendiagnosa penyakit pada paru-paru adalah dengan mendengarkan suara pernapasan pada paru-paru dengan stetoskop. Suara paru-paru yang dihasilkan pada beberapa kasus penyakit berbeda-beda sehingga terdapat pola tertentu yang bisa dikenali. Pola suara ini dapat digunakan untuk mengklasifikasikan jenis-jenis penyakit pada paru-paru. Masalah yang timbul adalah suara paru-paru menempati frekuensi yang cukup rendah sekitar 20-2000 Hz, amplitudo yang rendah, masalah kebisingan lingkungan, kepekaan telinga dan pola suara yang mirip antara jenis suara paru-paru yang satu dengan yang lain. Karena faktor-faktor tersebut diatas, kesalahan diagnosis biasa terjadi apabila prosedur auskultasi tidak dilakukan dengan benar. Pada penelitian ini dilakukan perbandingan klasifikasi suara paru-paru normal dan suara paru-paru abnormal untuk menentukan metode terbaik diantara dua metode yang akan digunakan. Suara paru-paru akan didekomposisi dengan menggunakan metode wavelet Daubechies2 level 5. Pada proses klasifikasi akan menggunakan metode klasifikasi Support Vector Machine One Against All, Support Vector Machine One Against One, dan klasifikasi jaringan syaraf tiruan PNN (Probabilistic Neural Network). Hasil penelitian ini adalah: akurasi rata-rata Support Vector Machine One against all sebesar 47,55%, maksimum akurasi sebesar 70%, dan rata-rata waktu komputasi selama 0.006 detik. Akurasi rata-rata Support Vector Machine One Against One sebesar 50.92%, maksimum akurasi sebesar 75%, dan rata-rata waktu komputasi selama 0.012 detik. Akurasi rata-rata Probabilistic Neural Network 70%, maksimum akurasi sebesar 70%, dan rata-rata waktu komputasi selama 0.313 detik.

Research paper thumbnail of Indonesian Vehicle Plate Recognition and Identification Based on Digital Image Processing and Artificial Neural Network

every vehicle has its own license which is given legally. In Indonesia, vehicle registration numb... more every vehicle has its own license which is given legally. In Indonesia, vehicle registration number is an important thing used by system such as: parking system, building security system and toll system. This research is doing a vehicle plate recognition and identification which will read the characters on the plate. The input comes from a real time video. There are six main processes in this system. In preprocessing, the system will enhance frame by using top hat and bottom hat transformation. While in detection, it will detect and crop the plate by using integral protection. After getting the plate position, an identification process will do by doing segmentation in each character. Then a feature extraction which combined with Artificial Neural Network is done to get the character of the vehicle registration number. The best Artificial Neural Network accuracy for character identification is obtained by using hexagonal topology and box distance with average 97,81 % accuracy.

Research paper thumbnail of Diaphysis Fracture on Tibia and Fibula Detection Based on Digital Image Processing and Scan Line Algorithm

IFMBE Proceedings, 2014

The X-Ray images of tibia and fibula are the important aiding for clinical diagnosis of fracture ... more The X-Ray images of tibia and fibula are the important aiding for clinical diagnosis of fracture because detection fracture performed by medical practice based on it. Under conditions of tired eyes, some medical practice miss fracture case. There are many previous research developed various methods to process X-Ray images and make system that can detect fracture automatically. In this paper, we proposes the simple system that can detect fracture on the tibia and fibula in two stages. The first stage image pre-processing using image enhancement, edge detection, filtering to remove noise and results the edge of image perfectly. The second stage is feature extraction using scan line algorithm that find maximum value of the difference distance to the nearest pixels from the right and left border margin of image. The maximum value of scan line is used as a threshold to classify normal cases or fracture cases. Accuracy is calculated based on images were tested right images for all images tested. Total images that used are 70 images, 30 normal images and 40 fracture images. Accuracy in this simulation reach 100% for normal images and 90% for fracture images, with time computation about 2.33 seconds.

Research paper thumbnail of Deteksi dan Klasifikasi Kelainan Jantung Berdasarkan Sinyal Elektrokardiograf Menggunakan Wavelet dan Jaringan Syaraf Tiruan Self Organizing Maps

Research paper thumbnail of End-To-End Deep Learning Architecture for Continuous Blood Pressure Estimation Using Attention Mechanism

Sensors

Blood pressure (BP) is a vital sign that provides fundamental health information regarding patien... more Blood pressure (BP) is a vital sign that provides fundamental health information regarding patients. Continuous BP monitoring is important for patients with hypertension. Various studies have proposed cuff-less BP monitoring methods using pulse transit time. We propose an end-to-end deep learning architecture using only raw signals without the process of extracting features to improve the BP estimation performance using the attention mechanism. The proposed model consisted of a convolutional neural network, a bidirectional gated recurrent unit, and an attention mechanism. The model was trained by a calibration-based method, using the data of each subject. The performance of the model was compared to the model that used each combination of the three signals, and the model with the attention mechanism showed better performance than other state-of-the-art methods, including conventional linear regression method using pulse transit time (PTT). A total of 15 subjects were recruited, and ...

Research paper thumbnail of Feature Analysis of Smart Shoe Sensors for Classification of Gait Patterns

Sensors

Gait analysis is commonly used to detect foot disorders and abnormalities such as supination, pro... more Gait analysis is commonly used to detect foot disorders and abnormalities such as supination, pronation, unstable left foot and unstable right foot. Early detection of these abnormalities could help us to correct the walking posture and avoid getting injuries. This paper presents extensive feature analyses on smart shoes sensor data, including pressure sensors, accelerometer and gyroscope signals, to obtain the optimum combination of the sensors for gait classification, which is crucial to implement a power-efficient mobile smart shoes system. In addition, we investigated the optimal length of data segmentation based on the gait cycle parameters, reduction of the feature dimensions and feature selection for the classification of the gait patterns. Benchmark tests among several machine learning algorithms were conducted using random forest, k-nearest neighbor (KNN), logistic regression and support vector machine (SVM) algorithms for the classification task. Our experiments demonstrat...

Research paper thumbnail of Deteksi Penyakit Covid-19 Berdasarkan Citra X-Ray Menggunakan Deep Residual Network

ELKOMIKA: Jurnal Teknik Energi Elektrik, Teknik Telekomunikasi, & Teknik Elektronika

ABSTRAKPenyakit Coronavirus-2019 atau Covid-19 telah menjadi pandemi global dan menjadi masalah u... more ABSTRAKPenyakit Coronavirus-2019 atau Covid-19 telah menjadi pandemi global dan menjadi masalah utama yang harus segera dikendalikan. Salah satu cara yang dapat dilakukan adalah memutus rantai penyebaran virus tersebut dengan melakukan deteksi dan melalukan karantina. Pencitraan X-Ray dapat dijadikan alternatif dalam mempelajari Covid-19. X-Ray dianggap mampu menggambarkan kondisi paru-paru pada pasien Covid-19 dan dapat menjadi alat bantu diagnosa klinis. Pada penelitian ini, kami mengusulkan pendekatan deep learning berbasis residual deep network untuk deteksi Covid-19 melalui citra chest X-Ray. Evaluasi yang dilakukan untuk mengetahui performa metode yang diusulkan berupa precision, recall, F1, dan accuracy. Hasil eksperimen menunjukkan bahwa usulan metode ini memberikan precision, recall, F1 dan accuracy masing-masing 0,98, 0,95, 0,97 dan 99%. Pada masa mendatang, studi ini diharapkan dapat divalidasi dan kemudian digunakan untuk melengkapi diagnosa klinis oleh dokter.Kata kunci...

Research paper thumbnail of Optimization of Deep Neural Networks for Heartrate Estimation from Face Video Stream to Implement Smart Health-City

The Journal of Korean Institute of Communications and Information Sciences

Research paper thumbnail of Multi-wavelet level comparison on compressive sensing for MRI image reconstruction

Bulletin of Electrical Engineering and Informatics

In this study, we proposed compressive sampling for MRI reconstruction based on sparse representa... more In this study, we proposed compressive sampling for MRI reconstruction based on sparse representation using multi-wavelet transformation. Comparing the performance of wavelet decomposition level, which are Level 1, Level 2, Level 3, and Level 4. We used gaussian random process to generate measurement matrix. The algorithm used to reconstruct the image is . The experimental results showed that the use of wavelet multi-level can generate higher compression ratio but requires a longer processing time. MRI reconstruction results based on the parameters of the peak signal to noise ratio (PSNR) and structural similarity index measure (SSIM) show that the higher the level of decomposition in wavelets, the value of both decreases.

Research paper thumbnail of DA-Capnet: Dual Attention Deep Learning Based on U-Net for Nailfold Capillary Segmentation

IEEE Access

Automatic nailfold capillary segmentation is a challenging task owing to noise and large variabil... more Automatic nailfold capillary segmentation is a challenging task owing to noise and large variabilities in images caused by insufficient focusing and low visibility of the capillaries. This task can be useful to detect and estimate the severity of autoimmune diseases of connective tissues or learning the status of white blood cells based on the cells' blood flow on the nailfold capillary. Previous studies have addressed this task using manual, semi-automated, and automated segmentation method. However, further improvement is still required. With the recent progress of deep learning on medical imaging, we herein propose dual attention deep learning based on U-Net for nailfold capillary segmentation, named DA-CapNet. Our DA-CapNet improves the U-Net architecture by integrating a dual attention module that can capture a better representation of feature maps from input images. Furthermore, DA-CapNet is compared with three baselines: adaptive Gaussian algorithm, SegNet, the original U-Net. We experimentally demonstrate that our proposed method outperforms these baselines.

Research paper thumbnail of Perancangan Aplikasi Pembaca Warna Untuk Penderita Buta Warna Berbasis Android

Jurnal Elektro dan Telekomunikasi Terapan

Penyandang cacat buta warna selalu mengalami kesulitan dalam hal membedakan warna dalam kehidupan... more Penyandang cacat buta warna selalu mengalami kesulitan dalam hal membedakan warna dalam kehidupan sehari-harinya, sehingga dibutuhkan alat bantu bagi mereka. Dalam penelitian ini dibuat sebuah algorithma yang dapat berfungsi untuk membedakan warna, algorithma tersebut diaplikasikan kedalam sebuah android. Aplikasi ini membantu dalam hal membedakan warna (merah, hijau dan biru) dengan pola image processing. Output dari aplikasi ini berupa text dan voice yang merupakan hasil deteksi dari input (captured image) secara langsung menggunakan smartphone android. Hasil yang diperoleh dari penelitian ini, yaitu tingkat akurasi total sistem 85% dengan waktu komputasi 6,67 detik dan jarak terjauh 20 cm.

Research paper thumbnail of Perancangan Dan Implementasi Tuner Gitar Otomatis Dengan Penggerak Motor Servo Berbasis Arduino

Jurnal Elektro dan Telekomunikasi Terapan

Gitar merupakan salah satu alat musik yang paling digemari, dan dapat dimainkan oleh semua orang.... more Gitar merupakan salah satu alat musik yang paling digemari, dan dapat dimainkan oleh semua orang. Saat ini masih banyak ditemukan pengguna gitar yang tidak bisa melakukan tuning senar gitar dan jika dilakukan manual juga membutuhkan waktu yang lama untuk mendapatkan hasil tuning yang akurat. Hal ini dapat diatasi dengan perkembangan teknologi Pengguna gitar sekarang ini dapat melakukan tuning senar lebih mudah dan lebih cepat dengan tingkat toleransi kesalahan frekuensi tuning ± 1 Hz dengan menggunakan tuner gitar otomatis.Perancangan alat tuner gitar otomatis ini menggunakan 5 komponen pembentuk alat yaitu selector switch untuk memilih frekuensi yang akan diatur, Op amp untuk menguatkan amplitudo gelombang suara agar diterima arduino dengan baik, Arduino untuk menginisialisasi frekuensi dan memberi sinyal ke motor servo, Motor servo untuk melakukan tuning senar gitar, dan LED sebagai indikator tuning.Hasil dari penelitian ini adalah keakuratan tuning untuk setiap senar didapatkan ...

Research paper thumbnail of Realisasi Pengendali Intensitas Cahaya Lampu Dengan Kontrol Suara Dan Google Android Speech Recognition Api

Jurnal Elektro dan Telekomunikasi Terapan

Dalam kehidupan sehari-hari setiap orang tidak dapat lepas dari cahaya untuk melakukan berbagai a... more Dalam kehidupan sehari-hari setiap orang tidak dapat lepas dari cahaya untuk melakukan berbagai aktivitas di ruangan. Namun cahaya di ruangan harus disesuaikan dengan kebutuhan aktivitas masing-masing untuk menghindari pemborosan energi listrik. Terkadang sebagian orang sering lupa dan malas untuk mematikan atau menyalakan lampu dalam keadaan tertentu karena saklar yang digunakan masih manual. Oleh karena itu dibutuhan saklar yang dapat dikendalikan dari jarak jauh. Penelitian ini merealisasikan sistem pengendali lampu jarak jauh berbasis suara yang didukung oleh google voice recognition engine dan menggunakan android , tidak hanya mematikan dan menghidupkan lampu, namun juga mengatur lampu dengan beberapa level intensitas cahaya. Arduino Uno R3 digunakan sebagai piranti pengendali dengan Bluetooth sebagai media komunikasi antara smartphone dan mikrokontroller. Berdasarkan pengujian sistem dan alat yang direalisasikan, untuk pengujian di dalam ruangan dengan penghalang, maksimal jan...

Research paper thumbnail of Routing Implementation Based-On Software Defined Network Using Ryu Controller and Openvswitch

Jurnal Teknologi, 2016

Open Flow is a standard protocol for differentiating forward function and control functions to fa... more Open Flow is a standard protocol for differentiating forward function and control functions to facilitate the management of big network of SDN. The research have been carried out before using the emulator SDN Mininet. However Mininet has many shortcomings, such as the performance of which is less than the maximum due to simulation. Then some researchers also use the Net-FPGA as device. This device is less suitable for small scale because the prices are quite expensive and programming is quite complicated. In this study, SDN implementation carried out using OpenvSwitch as forwarding function mounted on TP-Link that has modificated using openwrt as firmware and Raspberry Pi with Ryu SDN Controller as control functions. The result shows that routing static can be implemented on SDN Network which use Raspberry Pi with Ryu Controller as control function with average bandwith 536.0909 Mbits/sec and average uptime network is 10.45 second.

Research paper thumbnail of Pengenalan Plat Nomor Kendaraan Indonesia Berbasis Pengolahan Citra Digital Dan Jaringan Syaraf Tiruan Self-Organizing Map (Som)

Research paper thumbnail of Aplikasi Pendeteksi Penyakit Filariasis Berbasis Citra Darah

Research paper thumbnail of End-to-End Deep Learning Architecture for Continuous Blood Pressure Estimation Using Attention Mechanism

Sensors, Apr 20, 2020

Blood pressure (BP) is a vital sign that provides fundamental health information regarding patien... more Blood pressure (BP) is a vital sign that provides fundamental health information regarding patients. Continuous BP monitoring is important for patients with hypertension. Various studies have proposed cuff-less BP monitoring methods using pulse transit time. We propose an end-to-end deep learning architecture using only raw signals without the process of extracting features to improve the BP estimation performance using the attention mechanism. The proposed model consisted of a convolutional neural network, a bidirectional gated recurrent unit, and an attention mechanism. The model was trained by a calibration-based method, using the data of each subject. The performance of the model was compared to the model that used each combination of the three signals, and the model with the attention mechanism showed better performance than other state-of-the-art methods, including conventional linear regression method using pulse transit time (PTT). A total of 15 subjects were recruited, and electrocardiogram, ballistocardiogram, and photoplethysmogram levels were measured. The 95% confidence interval of the reference BP was [86.34, 143.74] and [51.28, 88.74] for systolic BP (SBP) and diastolic BP (DBP), respectively. The R 2 values were 0.52 and 0.49, and the mean-absolute-error values were 4.06 ± 4.04 and 3.33 ± 3.42 for SBP and DBP, respectively. In addition, the results complied with global standards. The results show the applicability of the proposed model as an analytical metric for BP estimation.

Research paper thumbnail of Automated White Blood Cell Counting in Nailfold Capillary Using Deep Learning Segmentation and Video Stabilization

Sensors

White blood cells (WBCs) are essential components of the immune system in the human body. Various... more White blood cells (WBCs) are essential components of the immune system in the human body. Various invasive and noninvasive methods to monitor the condition of the WBCs have been developed. Among them, a noninvasive method exploits an optical characteristic of WBCs in a nailfold capillary image, as they appear as visual gaps. This method is inexpensive and could possibly be implemented on a portable device. However, recent studies on this method use a manual or semimanual image segmentation, which depends on recognizable features and the intervention of experts, hindering its scalability and applicability. We address and solve this problem with proposing an automated method for detecting and counting WBCs that appear as visual gaps on nailfold capillary images. The proposed method consists of an automatic capillary segmentation method using deep learning, video stabilization, and WBC event detection algorithms. Performances of the three segmentation algorithms (manual, conventional, ...