Development and clinical implementation of SeedNet: A sliding-window convolutional neural network for radioactive seed identification in MRI-assisted radiosurgery (MARS)

2019 ◽  
Vol 81 (6) ◽  
pp. 3888-3900 ◽  
Author(s):  
Jeremiah W. Sanders ◽  
Steven J. Frank ◽  
Rajat J. Kudchadker ◽  
Teresa L. Bruno ◽  
Jingfei Ma
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Sliding Window ◽  
Clinical Implementation ◽  
Radioactive Seed ◽  
Seed Identification

scholarly journals Fast Wearable Sensor–Based Foot–Ground Contact Phase Classification Using a Convolutional Neural Network with Sliding-Window Label Overlapping

Sensors ◽  
2020 ◽  
Vol 20 (17) ◽  
pp. 4996 ◽  
Author(s):  
Haneul Jeon ◽  
Sang Lae Kim ◽  
Soyeon Kim ◽  
Donghun Lee
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Lower Limb ◽  
Sliding Window ◽  
Swing Phase ◽  
Training Dataset ◽  
Test Accuracy ◽  
Ground Contact ◽  
Contact Phase ◽  
Walking Gait

Classification of foot–ground contact phases, as well as the swing phase is essential in biomechanics domains where lower-limb motion analysis is required; this analysis is used for lower-limb rehabilitation, walking gait analysis and improvement, and exoskeleton motion capture. In this study, sliding-window label overlapping of time-series wearable motion data in training dataset acquisition is proposed to accurately detect foot–ground contact phases, which are composed of 3 sub-phases as well as the swing phase, at a frequency of 100 Hz with a convolutional neural network (CNN) architecture. We not only succeeded in developing a real-time CNN model for learning and obtaining a test accuracy of 99.8% or higher, but also confirmed that its validation accuracy was close to 85%.

Identifikasi Motif Jepara pada Ukiran dengan Memanfaatkan Convolutional Neural Network

2020 ◽  
Vol 9 (4) ◽  
pp. 403-413
Author(s):  
Sandhopi ◽  
Lukman Zaman P.C.S.W ◽  
Yosi Kristian
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Transfer Learning ◽  
Color Space ◽  
Sliding Window ◽  
Heat Maps

Semakin berkembang motif ukiran, semakin beragam bentuk dan variasinya. Hal ini menyulitkan dalam menentukan suatu ukiran bermotif Jepara. Pada makalah ini, metode transfer learning dengan FC yang dikembangkan dimanfaatkan untuk mengidentifikasi motif khas Jepara pada suatu ukiran. Dataset dibedakan menjadi tiga color space, yaitu LUV, RGB, dan YcrCb. Selain itu, sliding window, non-max suppression, dan heat maps dimanfaatkan untuk proses penelusuran area objek ukiran dan pengidentifikasian motif Jepara. Hasil pengujian dari semua bobot menunjukkan bahwa Xception pada klasifikasi motif Jepara memiliki nilai akurasi tertinggi, yaitu 0,95, 0,95, dan 0,94 untuk masing-masing dataset color space LUV, RGB, dan YCrCb. Namun, ketika semua bobot model tersebut diterapkan pada sistem identifikasi motif Jepara, ResNet50 mampu mengungguli semua jaringan dengan nilai persentase identifikasi motif sebesar 84%, 79%, dan 80%, untuk masing-masing color space LUV, RGB, dan YCrCb. Hasil ini membuktikan bahwa sistem mampu membantu dalam proses menentukan suatu ukiran, termasuk ke dalam ukiran Jepara atau bukan, dengan mengidentifikasi motif-motif khas Jepara yang terdapat dalam ukiran.

scholarly journals wUUNet: Advanced Fully Convolutional Neural Network for Multiclass Fire Segmentation

Symmetry ◽  
2021 ◽  
Vol 13 (1) ◽  
pp. 98
Author(s):  
Vladimir Sergeevich Bochkov ◽  
Liliya Yurievna Kataeva
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Input Data ◽  
Final Section ◽  
State Of The Art ◽  
Sliding Window ◽  
Input Image ◽  
Gaussian Mixtures ◽  
Flame Suppression ◽  
Segmentation Task

This article describes an AI-based solution to multiclass fire segmentation. The flame contours are divided into red, yellow, and orange areas. This separation is necessary to identify the hottest regions for flame suppression. Flame objects can have a wide variety of shapes (convex and non-convex). In that case, the segmentation task is more applicable than object detection because the center of the fire is much more accurate and reliable information than the center of the bounding box and, therefore, can be used by robotics systems for aiming. The UNet model is used as a baseline for the initial solution because it is the best open-source convolutional neural network. There is no available open dataset for multiclass fire segmentation. Hence, a custom dataset was developed and used in the current study, including 6250 samples from 36 videos. We compared the trained UNet models with several configurations of input data. The first comparison is shown between the calculation schemes of fitting the frame to one window and obtaining non-intersected areas of sliding window over the input image. Secondarily, we chose the best main metric of the loss function (soft Dice and Jaccard). We addressed the problem of detecting flame regions at the boundaries of non-intersected regions, and introduced new combinational methods of obtaining output signal based on weighted summarization and Gaussian mixtures of half-intersected areas as a solution. In the final section, we present UUNet-concatenative and wUUNet models that demonstrate significant improvements in accuracy and are considered to be state-of-the-art. All models use the original UNet-backbone at the encoder layers (i.e., VGG16) to demonstrate the superiority of the proposed architectures. The results can be applied to many robotic firefighting systems.

scholarly journals Recognition and Positioning of Container Lock Holes for Intelligent Handling Terminal Based on Convolutional Neural Network

2021 ◽  
Vol 38 (2) ◽  
pp. 467-472
Author(s):  
Xue Wang
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Local Binary Pattern ◽  
Sliding Window ◽  
Recognition Algorithm ◽  
Time Cost ◽  
Multi Scale ◽  
Container Transport ◽  
Automated Handling ◽  
Classifier Training

Container handling is a key link in container transport. In an automated handling terminal, the work efficiency directly depends on the time cost of the alignment between the spreader and the lock holes of the container. This paper attempts to improve the recognition and location of container lock holes with the aid of machine vision. Firstly, a lock hole recognition algorithm was designed based on local binary pattern (LBP) feature and classifier. After feature extraction and classifier training, multi-scale sliding window was used to recognize each lock hole. To realize real-time, accurate recognition of lock holes, the convolutional neural network (CNN) with improved threshold was incorporated to our algorithm. The tests on actual datasets show that our algorithm can effectively locate container lock holes.

scholarly journals Medical Image Classification Algorithm Based on Weight Initialization-Sliding Window Fusion Convolutional Neural Network

Complexity ◽  
2019 ◽  
Vol 2019 ◽  
pp. 1-15 ◽  
Author(s):  
Feng-Ping An
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
Convolutional Neural Network ◽  
Image Classification ◽  
Medical Image ◽  
Medical Images ◽  
Sliding Window ◽  
Weight Initialization ◽  
Medical Image Classification ◽  
Deep Learning Model

Due to the complexity of medical images, traditional medical image classification methods have been unable to meet actual application needs. In recent years, the rapid development of deep learning theory has provided a technical approach for solving medical image classification tasks. However, deep learning has the following problems in medical image classification. First, it is impossible to construct a deep learning model hierarchy for medical image properties; second, the network initialization weights of deep learning models are not well optimized. Therefore, this paper starts from the perspective of network optimization and improves the nonlinear modeling ability of the network through optimization methods. A new network weight initialization method is proposed, which alleviates the problem that existing deep learning model initialization is limited by the type of the nonlinear unit adopted and increases the potential of the neural network to handle different visual tasks. Moreover, through an in-depth study of the multicolumn convolutional neural network framework, this paper finds that the number of features and the convolution kernel size at different levels of the convolutional neural network are different. In contrast, the proposed method can construct different convolutional neural network models that adapt better to the characteristics of the medical images of interest and thus can better train the resulting heterogeneous multicolumn convolutional neural networks. Finally, using the adaptive sliding window fusion mechanism proposed in this paper, both methods jointly complete the classification task of medical images. Based on the above ideas, this paper proposes a medical classification algorithm based on a weight initialization/sliding window fusion for multilevel convolutional neural networks. The methods proposed in this study were applied to breast mass, brain tumor tissue, and medical image database classification experiments. The results show that the proposed method not only achieves a higher average accuracy than that of traditional machine learning and other deep learning methods but also is more stable and more robust.

scholarly journals Applications of neural networks for crack initiation and propagation monitoring in aircraft structures

2021 ◽  
pp. 99-103
Author(s):  
Надія Іванівна Бурау ◽  
Святослав Сергійович Юцкевич ◽  
Андрій Ігорович Компанець
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Crack Detection ◽  
Focal Length ◽  
Fatigue Cracks ◽  
Sliding Window ◽  
Detection Methods ◽  
Main Task ◽  
Learning Technology ◽  
Special Equipment

Timely detection of fatigue cracks on aircraft structural elements is the main task in damage tolerance principle approach. In this regard, much attention in aviation is paid to the methods of non-destructive testing which requires special equipment with the involvement of highly qualified personnel. Nowadays we can see that technologies that can learn to identify defects are preferred to simplify the gap process and minimize human factor errors. A self-learning technology is incorporated in the crack detection program. This makes it possible to increase the sensitivity of defects in the mode of the used technically false equipment. Unlike the detection methods of other machine learning detection systems, the system developed in this paper can also measure the cracks without the use of sophisticated sensors. However, the proposed system requires a photo-capturing device. Compared to similar visual systems, the developed system can work with very noisy images and detect cracks up to 0.3 mm. To do this, the webcam from the mid-range segment with 1920×1080 resolutions is used, that makes such technology easy to access. All modifications in the design of the camera scheme were associated with a change in the focal length, implemented by shifting the lens relative to the matrix. It allows the camera to focus on close distance less than 50 cm. For the fatigue tests compact specimens of duralumin alloy D16T with edge stress concentrator were used. The specimens were cycle tested by cantilever banding with stress ratio R=-1. Loading bogie apply force to specimens in direction normal to specimen surface. A loading value depends on the length of the loading crank and can be adjusted if needed.  To measure cracks in the processed images, a visual control program on a convolutional neural network and a sliding window algorithm were used. About 4,000 images were used to train the algorithm. The sliding window algorithm analyzes small images sequentially. One by one, image regions were selected and monitored for cracks using a convolutional neural network. Areas with detected cracks are memorized by the sliding window algorithm.

CONVOLUTIONAL NEURAL NETWORK BASED ALGORITHM FOR CECUM ACHIEVEMENT CONFIRMATION

2020 ◽  
Author(s):  
S Kashin ◽  
D Zavyalov ◽  
A Rusakov ◽  
V Khryashchev ◽  
A Lebedev
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network
Sign in / Sign up

Export Citation Format

Share Document