ON SELECTING IMAGES FROM AN UNAIMED VIDEO STREAM FOR PHOTOGRAMMETRIC MODELLING

Abstract. In this paper, we illustrate how convolutional neural networks and voxel-based processing together with voxel visualizations can be utilized for the selection of unaimed images for a photogrammetric image block. Our research included the detection of an ear from images with a convolutional neural network, computation of image orientations with a structure-from-motion algorithm, visualization of camera locations in a voxel representation to detect the goodness of the imaging geometry, rejection of unnecessary images with an XYZ buffer, the creation of 3D models in two different example cases, and the comparison of resulting 3D models. Two test data sets were taken of an ear with the video recorder of a mobile phone. In the first test case, a special emphasis was taken to ensure good imaging geometry. On the contrary, in the second test case the trajectory was limited to approximately horizontal movement, leading to poor imaging geometry. A convolutional neural network together with an XYZ buffer managed to select a useful set of images for the photogrammetric 3D measuring phase. The voxel representation well illustrated the imaging geometry and has potential for early detection where data is suitable for photogrammetric modelling. The comparison of 3D models revealed that the model from poor imaging geometry was noisy and flattened. The results emphasize the importance of good imaging geometry.

Download Full-text

Spectral Convolution Feature-Based SPD Matrix Representation for Signal Detection Using a Deep Neural Network

Entropy ◽

10.3390/e22090949 ◽

2020 ◽

Vol 22 (9) ◽

pp. 949

Author(s):

Jiangyi Wang ◽

Min Liu ◽

Xinwu Zeng ◽

Xiaoqiang Hua

Keyword(s):

Neural Network ◽

Signal Detection ◽

Convolutional Neural Network ◽

Deep Neural Network ◽

Detection Method ◽

Learning Algorithm ◽

Simulated Data ◽

Data Sets ◽

Feature Maps ◽

Simulated Data Sets

Convolutional neural networks have powerful performances in many visual tasks because of their hierarchical structures and powerful feature extraction capabilities. SPD (symmetric positive definition) matrix is paid attention to in visual classification, because it has excellent ability to learn proper statistical representation and distinguish samples with different information. In this paper, a deep neural network signal detection method based on spectral convolution features is proposed. In this method, local features extracted from convolutional neural network are used to construct the SPD matrix, and a deep learning algorithm for the SPD matrix is used to detect target signals. Feature maps extracted by two kinds of convolutional neural network models are applied in this study. Based on this method, signal detection has become a binary classification problem of signals in samples. In order to prove the availability and superiority of this method, simulated and semi-physical simulated data sets are used. The results show that, under low SCR (signal-to-clutter ratio), compared with the spectral signal detection method based on the deep neural network, this method can obtain a gain of 0.5–2 dB on simulated data sets and semi-physical simulated data sets.

Download Full-text

AI INFLUENCE IN COVID-19 DETECTION

Journal of University of Shanghai for Science and Technology ◽

10.51201/jusst/21/07235 ◽

2021 ◽

Vol 23 (07) ◽

pp. 1116-1120

Author(s):

Cijil Benny ◽

Keyword(s):

Neural Network ◽

Statistical Analysis ◽

Convolutional Neural Network ◽

Case Reports ◽

Ct Scans ◽

Data Sets ◽

X Rays ◽

Medical Field ◽

Primary Key ◽

Made In

This paper is on analyzing the feasibility of AI studies and the involvement of AI in COVID interrelated treatments. In all, several procedures were reviewed and studied. It was on point. The best-analyzing methods on the studies were Susceptible Infected Recovered and Susceptible Exposed Infected Removed respectively. Whereas the implementation of AI is mostly done in X-rays and CT- Scans with the help of a Convolutional Neural Network. To accomplish the paper several data sets are used. They include medical and case reports, medical strategies, and persons respectively. Approaches are being done through shared statistical analysis based on these reports. Considerably the acceptance COVID is being shared and it is also reachable. Furthermore, much regulation is needed for handling this pandemic since it is a threat to global society. And many more discoveries shall be made in the medical field that uses AI as a primary key source.

Download Full-text

HEp-2 CELL CLASSIFICATION BY ADAPTIVE CONVOLUTIONAL LAYER BASED CONVOLUTIONAL NEURAL NETWORK

Biomedical Engineering Applications Basis and Communications ◽

10.4015/s1016237219500443 ◽

2019 ◽

Vol 31 (06) ◽

pp. 1950044

Author(s):

C. C. Manju ◽

M. Victor Jose

Keyword(s):

Neural Network ◽

Autoimmune Diseases ◽

Convolutional Neural Network ◽

Research Work ◽

Test Case ◽

Morphological Operations ◽

Cell Classification ◽

Proposed Model ◽

Sensitivity Specificity ◽

Opening And Closing

Objective: The antinuclear antibodies (ANA) that present in the human serum have a link with various autoimmune diseases. Human Epithelial type-2 (HEp-2) cells acts as a substance in the Indirect Immuno fluorescence (IIF) test for diagnosing these autoimmune diseases. In recent times, the computer-aided diagnosis of autoimmune diseases by the HEp-2 cell classification has drawn more interest. Though, they often pose limitations like large intra-class and small inter-class variations. Hence, various efforts have been performed to automate the procedure of HEp-2 cell classification. To overcome these problems, this research work intends to propose a new HEp-2 classification process. Materials and Methods: This is regulated by integrating two processes, namely, segmentation and classification. Initially, the segmentation of the HEp-2 cells is carried out by deploying the morphological operations. In this paper, two morphology operations are deployed called opening and closing. Further, the classification process is exploited by proposing a modified Convolutional Neural Network (CNN). The main objective is to classify the HEp-2 cells effectively (Centromere, Golgi, Homogeneous, Nucleolar, NuMem, and Speckled) and is made by exploiting the optimization concept. This is implanted by developing a new algorithm called Distance Sorting Lion Algorithm (DSLA), which selects the optimal convolutional layer in CNN. Results: Through the performance analysis, the performance of the proposed model for test case 1 at learning percentage 60 is 3.84%, 1.79%, 6.22%, 1.69%, and 5.53% better than PSO, FF, GWO, WOA, and LA, respectively. At 80, the performance of the proposed model is 5.77%, 6.46%, 3.95%, 3.24%, and 5.55% better from PSO, FF, GWO, WOA, and LA, respectively. Hence, the performance of the proposed work is proved over other models under different measures. Conclusion: Finally, the performance is evaluated by comparing it with the other conventional algorithms in terms of accuracy, sensitivity, specificity, precision, FPR, FNR, NPV, MCC, F1-Score and FDR, and proves the efficacy of the proposed model.

Download Full-text

An Efficient 3D Model Retrieval Method Based on Convolutional Neural Network

Complexity ◽

10.1155/2020/9050459 ◽

2020 ◽

Vol 2020 ◽

pp. 1-14

Author(s):

Bo Ding ◽

Lei Tang ◽

Yong-jun He

Keyword(s):

Neural Network ◽

Convolutional Neural Network ◽

3D Model ◽

3D Models ◽

3D Model Retrieval ◽

Similarity Matching ◽

Retrieval Method ◽

Model Classification ◽

Model Retrieval ◽

Index Building

Recently, 3D model retrieval based on views has become a research hotspot. In this method, 3D models are represented as a collection of 2D projective views, which allows deep learning techniques to be used for 3D model classification and retrieval. However, current methods need improvements in both accuracy and efficiency. To solve these problems, we propose a new 3D model retrieval method, which includes index building and model retrieval. In the index building stage, 3D models in library are projected to generate a large number of views, and then representative views are selected and input into a well-learned convolutional neural network (CNN) to extract features. Next, the features are organized according to their labels to build indexes. In this stage, the views used for representing 3D models are reduced substantially on the premise of keeping enough information of 3D models. This method reduces the number of similarity matching by 87.8%. In retrieval, the 2D views of the input model are classified into a category with the CNN and voting algorithm, and then only the features of one category rather than all categories are chosen to perform similarity matching. In this way, the searching space for retrieval is reduced. In addition, the number of used views for retrieval is gradually increased. Once there is enough evidence to determine a 3D model, the retrieval process will be terminated ahead of time. The variable view matching method further reduces the number of similarity matching by 21.4%. Experiments on the rigid 3D model datasets ModelNet10 and ModelNet40 and the nonrigid 3D model dataset McGill10 show that the proposed method has achieved retrieval accuracy rates of 94%, 92%, and 100%, respectively.

Download Full-text

Computer-Aided Bacillus Detection in Whole-Slide Pathological Images Using a Deep Convolutional Neural Network

Applied Sciences ◽

10.3390/app10124059 ◽

2020 ◽

Vol 10 (12) ◽

pp. 4059

Author(s):

Chung-Ming Lo ◽

Yu-Hung Wu ◽

Yu-Chuan (Jack) Li ◽

Chieh-Chi Lee

Keyword(s):

Neural Network ◽

Convolutional Neural Network ◽

Deep Convolutional Neural Network ◽

Color Information ◽

Training Set ◽

Image Block ◽

Negative Image ◽

Cad System ◽

Computer Aided ◽

Whole Slide Images

Mycobacterial infections continue to greatly affect global health and result in challenging histopathological examinations using digital whole-slide images (WSIs), histopathological methods could be made more convenient. However, screening for stained bacilli is a highly laborious task for pathologists due to the microscopic and inconsistent appearance of bacilli. This study proposed a computer-aided detection (CAD) system based on deep learning to automatically detect acid-fast stained mycobacteria. A total of 613 bacillus-positive image blocks and 1202 negative image blocks were cropped from WSIs (at approximately 20 × 20 pixels) and divided into training and testing samples of bacillus images. After randomly selecting 80% of the samples as the training set and the remaining 20% of samples as the testing set, a transfer learning mechanism based on a deep convolutional neural network (DCNN) was applied with a pretrained AlexNet to the target bacillus image blocks. The transferred DCNN model generated the probability that each image block contained a bacillus. A probability higher than 0.5 was regarded as positive for a bacillus. Consequently, the DCNN model achieved an accuracy of 95.3%, a sensitivity of 93.5%, and a specificity of 96.3%. For samples without color information, the performances were an accuracy of 73.8%, a sensitivity of 70.7%, and a specificity of 75.4%. The proposed DCNN model successfully distinguished bacilli from other tissues with promising accuracy. Meanwhile, the contribution of color information was revealed. This information will be helpful for pathologists to establish a more efficient diagnostic procedure.

Download Full-text

Diagnosis of left ventricular hypertrophybased on convolutional neural network

10.21203/rs.2.19295/v1 ◽

2019 ◽

Author(s):

Zini Jian ◽

Xianpei Wang ◽

Jingzhe Zhang ◽

Xinyu Wang ◽

Youbin Deng

Keyword(s):

Neural Network ◽

Left Ventricle ◽

Convolutional Neural Network ◽

Posterior Wall ◽

Video Stream ◽

Detection Algorithm ◽

Left Ventricular ◽

Diagnostic Significance ◽

Design Algorithm ◽

Actual Measurement

Abstract Background: Clinically, doctors obtain the left ventricular posterior wall thickness (LVPWT) mainly by observing ultrasonic echocardiographic video stream to capture a single frame of images with diagnostic significance, and then mark two key points on both sides of the posterior wall of the left ventricle with their own experience for computer measurement. In the actual measurement, the doctor's selection point is subjective, which is not only time-consuming and laborious, but also difficult to accurately locate the edge, which will bring errors to the measurement results. Methods: In this paper, a convolutional neural network model of left ventricular posterior wall positioning was built under the TensorFlow framework, and the target region images were obtained after the positioning results were processed by non-local mean filtering and opening operation. Then the edge detection algorithm based on threshold segmentation is used. After the contour was extracted by adjusting the segmentation threshold through prior analysis and the OTSU algorithm, the design algorithm completed the computer selection point measurement of the thickness of the posterior wall of the left ventricle. Results: The proposed method can effectively extract the left ventricular posterior wall contour and measure its thickness. The experimental results show that the relative error between the measurement result and the hospital measurement value is less than 15%, which is less than 20% of the acceptable repeatability error in clinical practice. Conclusions: Therefore, the method proposed in this paper not only has the advantage of less manual intervention, but also can reduce the workload of doctors.

Download Full-text

A New Classification Method in Ultrasound Images of Benign and Malignant Thyroid Nodules Based on Transfer Learning and Deep Convolutional Neural Network

Complexity ◽

10.1155/2021/6296811 ◽

2021 ◽

Vol 2021 ◽

pp. 1-9 ◽

Cited By ~ 1

Author(s):

Weibin Chen ◽

Zhiyang Gu ◽

Zhimin Liu ◽

Yaoyao Fu ◽

Zhipeng Ye ◽

...

Keyword(s):

Neural Network ◽

Convolutional Neural Network ◽

Transfer Learning ◽

Thyroid Nodules ◽

Ultrasound Image ◽

Image Data ◽

Data Sets ◽

Thyroid Ultrasound ◽

Joint Training ◽

Secondary Transfer

Thyroid nodule is a clinical disorder with a high incidence rate, with large number of cases being detected every year globally. Early analysis of a benign or malignant thyroid nodule using ultrasound imaging is of great importance in the diagnosis of thyroid cancer. Although the b-mode ultrasound can be used to find the presence of a nodule in the thyroid, there is no existing method for an accurate and automatic diagnosis of the ultrasound image. In this pursuit, the present study envisaged the development of an ultrasound diagnosis method for the accurate and efficient identification of thyroid nodules, based on transfer learning and deep convolutional neural network. Initially, the Total Variation- (TV-) based self-adaptive image restoration method was adopted to preprocess the thyroid ultrasound image and remove the boarder and marks. With data augmentation as a training set, transfer learning with the trained GoogLeNet convolutional neural network was performed to extract image features. Finally, joint training and secondary transfer learning were performed to improve the classification accuracy, based on the thyroid images from open source data sets and the thyroid images collected from local hospitals. The GoogLeNet model was established for the experiments on thyroid ultrasound image data sets. Compared with the network established with LeNet5, VGG16, GoogLeNet, and GoogLeNet (Improved), the results showed that using GoogLeNet (Improved) model enhanced the accuracy for the nodule classification. The joint training of different data sets and the secondary transfer learning further improved its accuracy. The results of experiments on the medical image data sets of various types of diseased and normal thyroids showed that the accuracy rate of classification and diagnosis of this method was 96.04%, with a significant clinical application value.

Download Full-text

Accurate Positioning of License Plate in Video Stream Based on Concatenated Convolutional Neural Network

Review of Computer Engineer Studies ◽

10.18280/rces.060203 ◽

2019 ◽

Vol 6 (2) ◽

pp. 44-49

Author(s):

Pan Ding ◽

Hua Sun ◽

Chuping Xiong ◽

Yao Li

Keyword(s):

Neural Network ◽

Convolutional Neural Network ◽

Video Stream ◽

License Plate

Download Full-text

Introspective analysis of convolutional neural networks for improving discrimination performance and feature visualisation

PeerJ Computer Science ◽

10.7717/peerj-cs.497 ◽

2021 ◽

Vol 7 ◽

pp. e497

Author(s):

Shakeel Shafiq ◽

Tayyaba Azim

Keyword(s):

Neural Network ◽

Neural Networks ◽

Convolutional Neural Network ◽

Convolutional Neural Networks ◽

Discrimination Performance ◽

Input Image ◽

Data Sets ◽

Discrimination Power ◽

Level Information ◽

Fully Connected

Deep neural networks have been widely explored and utilised as a useful tool for feature extraction in computer vision and machine learning. It is often observed that the last fully connected (FC) layers of convolutional neural network possess higher discrimination power as compared to the convolutional and maxpooling layers whose goal is to preserve local and low-level information of the input image and down sample it to avoid overfitting. Inspired from the functionality of local binary pattern (LBP) operator, this paper proposes to induce discrimination into the mid layers of convolutional neural network by introducing a discriminatively boosted alternative to pooling (DBAP) layer that has shown to serve as a favourable replacement of early maxpooling layer in a convolutional neural network (CNN). A thorough research of the related works show that the proposed change in the neural architecture is novel and has not been proposed before to bring enhanced discrimination and feature visualisation power achieved from the mid layer features. The empirical results reveal that the introduction of DBAP layer in popular neural architectures such as AlexNet and LeNet produces competitive classification results in comparison to their baseline models as well as other ultra-deep models on several benchmark data sets. In addition, better visualisation of intermediate features can allow one to seek understanding and interpretation of black box behaviour of convolutional neural networks, used widely by the research community.

Download Full-text

Core looseness fault identification model based on Mel spectrogram-CNN

Journal of Physics Conference Series ◽

10.1088/1742-6596/2137/1/012060 ◽

2021 ◽

Vol 2137 (1) ◽

pp. 012060

Author(s):

Ping He ◽

Yong Li ◽

Shoulong Chen ◽

Hoghua Xu ◽

Lei Zhu ◽

...

Keyword(s):

Neural Network ◽

Convolutional Neural Network ◽

Convolution Neural Network ◽

Data Sets ◽

Data Set ◽

Recognition Model ◽

Model Based ◽

Fault Recognition ◽

Transformer Core ◽

Identification Model

Abstract In order to realize transformer voiceprint recognition, a transformer voiceprint recognition model based on Mel spectrum convolution neural network is proposed. Firstly, the transformer core looseness fault is simulated by setting different preloads, and the sound signals under different preloads are collected; Secondly, the sound signal is converted into a spectrogram that can be trained by convolutional neural network, and then the dimension is reduced by Mel filter bank to draw Mel spectrogram, which can generate spectrogram data sets under different preloads in batch; Finally, the data set is introduced into convolutional neural network for training, and the transformer voiceprint fault recognition model is obtained. The results show that the training accuracy of the proposed Mel spectrum convolution neural network transformer identification model is 99.91%, which can well identify the core loosening faults.

Download Full-text