25 research outputs found
Metode Baru Semi Otomatis Berbasis Active Shape Model Untuk Penentuan Tingkat Keparahan Osteoarthritis Lutut
No full textOsteoarthritis (OA) merupakan penyakit arthritis yang sering ditemukan di masyarakat. Beberapa penelitian telah menetapkan bahwa Osteoarthritis (OA) ditemukan pada orang berusia 40-60 tahun dengan prevalensi 15,5% pada pria dan 12,7% pada wanita. Penyakit ini secara bertahap dapat menjadi kronis jika tidak dikontrol secara klinis dengan tepat. Gejala OA ditandai oleh nyeri sendi dan gangguan gerakan karena kerusakan tulang rawan. OA sering terjadi pada persendian lutut, pinggul, tulang belakang, dan kaki. OA dapat terjadi dengan etiologi yang berbeda-beda, namun mengakibatkan kelainan biologis, morfologis dan luaran klinis yang sama. Sehingga, studi mengenai OA dapat dilakukan secara multidisiplin. Identifikasi awal OA memainkan peran penting untuk pengambilan keputusan klinis, mendeteksi keparahan OA dan pengobatan yang tepat. Penyelarasan tulang femur-tibia merupakan faktor risiko utama untuk kejadian dan perkembangan knee osteoarthritis (KOA). Selain itu, untuk menentukan tingkat keparahan OA lutut dapat dilakukan dengan mengklasifikasikan tingkat keparahan berdasarkan KL grade dan Femur Tibia Angle (FTA). Pada penelitian ini, dilakukan penentuan tingkat keparahan OA lutut menggunakan beberapa metode hibrida baru pada citra radiografi (x-ray) yang diperoleh dari Osteoarthritis Initiative (OAI) database. Metode hibrida baru yang dimaksud adalah Structural Two Dimentional Principal Component Analysis (S2DPCA) + Support Vector Machine (SVM) dan Convolutional Neural Network (CNN) + Long Short Term Memory (LSTM) untuk penentuan tingkat keparahan OA lutut berdasarkan KL Grade serta Active Shape Model (ASM) untuk penentuan sudut femur dan tibia (FTA) secara semi otomatis. Hasil ujicoba pada metode S2DPCA + SVM menggunakan 80 citra x-ray lutut dengan ujicoba berdasarkan four-fold cross validation diperoleh hasil maksimal sebesar 94,33% untuk grade 0 dengan kernel Gaussian. Sedangkan untuk metode CNN + LSTM menggunakan 1530 citra x-ray lutut (1055 citra pelatihan dan 475 citra pengujian) dengan ujicoba berdasarkan three-fold cross validation diperoleh hasil maksimal sebesar 75,28% untuk rata-rata dari grade 0 – 4. Sementara itu, untuk metode ASM pada penentuan FTA yang dikembangkan (disebut TRIMA-FTA) dengan menggunakan 60 citra x-ray lutut (10 citra pelatihan dan 50 citra pengujian) diperoleh perbedaan hasil pengukuran antara FTA OAI dan FTA yang cukup kecil, yaitu rata-rata di bawah 0,810 untuk FTA kanan dan di bawah 0,770 untuk FTA kiri. Hasil-hasil tersebut menunjukkan bahwa pendekatan-pendekatan baru yang dikembangkan secara klinis cocok untuk penentuan tingkat keparahan OA lutut berdasarkan KL grade dan pengukuran FTA secara semi otomatis.
==================================================================================================================================
Osteoarthritis (OA) is an arthritis diseases that was found commonly in society. Numerous investigations have established that Osteoarthritis (OA) were found in people aged 40-60 years with a prevalence of 15.5% in men and 12.7% in women. This disease gradually became a chronic if not correctly clinical controlled. OA’s symptom is characterized by joint pain and movement disorders due to damage of the cartilage. OA often occurs in the joints of knees, hips, spines and feets. OA can be occur with different etiologies, but the results could be same in biological, morphological and clinical outcomes. Thus, studies on OA can be done in a multidisciplinary science. Early identification of OA plays an important role to improve clinical decision making, to monitor disease progress and appropriate treatment. Some researchers have conducted studies on the severity of knee OA based on Kellgren Lawrence (KL) Grade, Joint Space Width (JSW), osteophyte formation, as well as the angle between the femur bone and the tibia / Femur Tibia Angle (FTA). Furthermore, alignment of the femur-tibia bone is a major risk factor for the occurrence and development of knee osteoarthritis (KOA). In addition, to determine the severity of knee OA can be done by classifying the severity based on KL grade and Femur Tibia Angle (FTA). In this study, the severity of knee OA was determined using several new hybrid methods on radiographic (x-ray) images obtained from the Osteoarthritis Initiative (OAI) database. The new hybrid method is Structural Two Dimentional Principal Component Analysis (S2DPCA) + Support Vector Machine (SVM) and Convolutional Neural Network (CNN) + Long Short Term Memory (LSTM) for determining the severity of knee OA based on KL Grade and Active Shape Model (ASM) for semi-automatic determination of the angle of the femur and tibia (FTA). The results of trials on the S2DPCA + SVM method using 80 knee x-ray images with trials based on four-fold cross validation obtained maximum results of 94.33% for grade 0 with Gaussian kernel. Whereas the CNN + LSTM method uses 1530 knee x ray images (1055 training images and 475 test images) with trials based on three-fold cross validation, the maximum results are 75.28% for an average of grades 0 - 4. Meanwhile , for the ASM method in determining the developed FTA (called TRIMA FTA) using 60 knee x-ray images (10 training images and 50 test images), the difference in measurement results between FTA OAI and FTA is quite small, namely the average in below 0.810 for the right FTA and below 0.770 for the left FTA. These results indicate that new clinically developed approaches are suitable for determining the severity of knee OA based on KL grade and semi-automatic FTA measurements
SEGMENTASI OBYEK PADA CITRA DIGITAL MENGGUNAKAN METODE OTSU THRESHOLDING
Get PDFDigital image has size and object in the form of foreground and background. To separate it, it is necessary to be conducted the image segmentation process. Otsu thresholding method is one of image segmentation method. In this research is divided into five processes, which are input image, pre-processing, segmentation, cleaning, and accuracy calculation. First process was input color images which consists of multiple objects. Second process was conversion from color image to grayscale image. Third process was automatically calculated threshold value using Otsu thresholding method, followed by binary image transformation. The fourth process, the result of third process is changed into negative image as the segmentation results, noise removal with a threshold value of 150, and morphology. The last accuracy calculation is conducted to measure proposed segmentation method performance. The experimental result have been compared to the image of Ground Truth as the direct user observation to calculate accuracy. To examine the proposed method, Weizmann Segmentation Database is used as data set. It conconsist of 30 color images. The experimental results show that 93.33% accuracy were achieved
Deep Neural Networks for Automatic Classification of Knee Osteoarthritis Severity Based on X-ray Images
Get PDFKnee Osteoarthritis (KOA) is a type of chronic disease that commonly occurs in older, obese citizens and those with a sedentary lifestyle. This disease causes damage to knee cartilage and causes pain so that the patient's activity is reduced. Radiologists classify the KOA severity based on Joint Space Narrowing (JSN) and the presence or absence of osteophytes into five stages from healthyknee (stage 0) to the worst damage (stage 4). We developed a methodology that aims to accelerate the classification of KOA severity based on information obtained from X-ray images and to reduce the subjectivity of radiologists. This paper describes an automated KOA diagnostic model using hyper-parameter Deep Convolutional Neural Networks (DCNN). Based on our experimental result, it shows the accuracy of the proposed method outperforms other KOA severity classification algorithms, which alsodiscussed deep learning, namely 77.24%. This value is the average result of the accuracy of each fold from each stage of the KOA severity level where we use three-folds cross validation as a method of evaluating system performance. Thus computationally, this method is efficient in automatic diagnosis and has the potential to be a clinician application aid to specify the KOA severity
SEGMENTASI OBYEK PADA CITRA DIGITAL MENGGUNAKAN METODE OTSU THRESHOLDING
No full textDigital image has size and object in the form of foreground and background. To separate it, it is necessary to be conducted the image segmentation process. Otsu thresholding method is one of image segmentation method. In this research is divided into five processes, which are input image, pre-processing, segmentation, cleaning, and accuracy calculation. First process was input color images which consists of multiple objects. Second process was conversion from color image to grayscale image. Third process was automatically calculated threshold value using Otsu thresholding method, followed by binary image transformation. The fourth process, the result of third process is changed into negative image as the segmentation results, noise removal with a threshold value of 150, and morphology. The last accuracy calculation is conducted to measure proposed segmentation method performance. The experimental result have been compared to the image of Ground Truth as the direct user observation to calculate accuracy. To examine the proposed method, Weizmann Segmentation Database is used as data set. It conconsist of 30 color images. The experimental results show that 93.33% accuracy were achieved
Classification of Corn Seed Quality Using Convolutional Neural Network with Region Proposal and Data Augmentation
Get PDFCorn is a commodity in agriculture and essential to human food and animal feed. All components of corn can be utilized and accommodated for the benefit of humans. One of the supporting components is the quality of corn seeds, where specific sources have physiological properties to survive. The problem is how to get information on the quality of corn seeds at agricultural locations and get information through direct visual observations. This research tries to find a solution for classifying corn kernels with high accuracy using a convolutional neural network. It is because in-depth training is used in deep learning. The problem with convolutional neural networks is that the training process takes a long time, depending on the number of layers in the architecture. The research contribution is adding Convex Hull. This method looks for edge points on an object and forms a polygon that encloses that point. It helps increase focus on the convolution multiplication process by removing images on the background. The 34-layer architecture maintains feature maps and uses dropout layers to save computation time. The dataset used is primary data. There are six classes, AR21, Pioner_P35, BISI_18, NK212, Pertiwi, and Betras1—data augmentation techniques to overcome data limitations so that overfitting does not occur. The results of the classification of corn kernels obtained a model with an average accuracy of 99.33%, 99.33% precision, 99.33% recall, and 99.36% F-1 score. The computational training time to obtain the model was 2 minutes 30 seconds. The average error value for MSE is 0.0125, RMSE is 0.118, and MAE is 0.0108. The experimental data testing process has an accuracy ranging from 77% -99%. In conclusion, using the proposal area can improve accuracy by about 0.3% because the focused object helps the convolution process
PENGENALAN POLA SENYUM MENGGUNAKAN SELF ORGANIZING MAPS (SOM) BERBASIS EKSTRAKSI FITUR TWO-DIMENSIONAL PRINCIPAL COMPONENT ANALYSIS (2DPCA)
No full textPengenalan pola senyum merupakan bagian dari pattern recognition dan telah banyak dikembangkan. Penelitian ini bertujuan melakukan pengenalan pola senyum menjadi lima macam klasifikasi yaitu: senyum manis, senyum mulut tertutup, senyum mulut terbuka, senyum mengejek, senyum yang dipaksakan. Dua hal yang menjadi masalah utama pada identifikasi pola senyum adalah proses ekstraksi fitur dari sampel pola senyum yang ada dan juga teknik klasifikasi yang digunakan untuk mengklasifikasikan pola senyum yang ingin dikenali berdasarkan fitur-fitur yang telah dipilih. Pada penelitian ini dilakukan proses ekstraksi fitur menggunakan algoritma Two-Dimensional Principal Component Analysis (2DPCA), sedangkan untuk proses klasifikasi menggunakan algoritma Self Organizing Maps (SOM). Berdasarkan hasil ujicoba menggunakan 250 data wajah tersenyum, rata-rata akurasi pengenalan pola senyum tertinggi yaitu sebesar 93.36% untuk 30 sampel masing-masing pola senyum
