- 标题
- 摘要
- 关键词
- 实验方案
- 产品
过滤筛选
- 2018
- Conditional Random Fields (CRF)
- Convolutional Neural Network (CNN)
- Fine Classification
- Airborne hyperspectral
- Optoelectronic Information Science and Engineering
- Wuhan University
- Central South University
- Hubei University
-
[IEEE 2018 IEEE 3rd International Conference on Image, Vision and Computing (ICIVC) - Chongqing (2018.6.27-2018.6.29)] 2018 IEEE 3rd International Conference on Image, Vision and Computing (ICIVC) - Low Resolution Cell Image Edge Segmentation Based on Convolutional Neural Network
摘要： An low-resolution cell images captured by a lens-free imaging system is presented in this paper. The resolution of this cell images is impacted by the low and experimental cell segmentation methods to solve the original cell images is not robust and sensitive to noise. So based on the convolutional neural network, an optimized CSnet method is proposed in this paper for automatically segmenting cell. In the proposed method, the produced data set will be sent into the convolutional neural network firstly for training to obtain an optimized convolution neural network segmentation model. And then, the pre-divided images acquired by the lens-free imaging system are loaded into the segmentation model to get the segmentation images. Finally, our proposed method in this paper is tested in a neural network framework built in keras. The experimental results show that the accuracy of our proposed method can reach about 96%. At the same time, it also can implement batch segmentation automatically and make the problem of heavy task for segmentation better.
关键词： convolutional neural network,cell segmentation,Lensfree imaging,microfluidic chip
更新于2025-09-10 09:29:36
-
[Advances in Intelligent Systems and Computing] Soft Computing for Problem Solving Volume 817 (SocProS 2017, Volume 2) || Image Compression Using Neural Network for Biomedical Applications
摘要： As images are of large size and require huge bandwidth and large storage space, an effective compression algorithm is essential. Hence in this paper, feedforward backpropagation neural network with the multilayer perception using resilient backpropagation (RP) algorithm is used with the objective to develop an image compression in the field of biomedical sciences. With the concept of neural network, data compression can be achieved by producing an internal data representation. This network is an application of backpropagation that takes huge content of data as input, compresses it while storing or transmitting, and decompresses the compressed data whenever required. The training algorithm and development architecture give less distortion and considerable compression ratio and also keep up the capability of hypothesizing and are becoming important. The efficiency of the RP is evaluated on x-ray image of rib cage and has given better results of the various performance metrics when compared to the other algorithms.
关键词： Artificial neural network,Backpropagation neural network,Gradient descent algorithm (GD),Resilient backpropagation algorithm (RP),Image compression
更新于2025-09-09 09:28:46
-
[Lecture Notes in Computer Science] Neural Information Processing Volume 11307 (25th International Conference, ICONIP 2018, Siem Reap, Cambodia, December 13–16, 2018, Proceedings, Part VII) || Hopfield Neural Network with Double-Layer Amorphous Metal-Oxide Semiconductor Thin-Film Devices as Crosspoint-Type Synapse Elements and Working Confirmation of Letter Recognition
摘要： Arti?cial intelligences are essential concepts in smart societies, and neural networks are typical schemes that imitate human brains. However, the neural networks are conventionally realized using complicated software and high-performance hardware, and the machine size and power consumption are huge. On the other hand, neuromorphic systems are composed solely of optimized hardware, and the machine size and power consumption can be reduced. Therefore, we are investigating neuromorphic systems especially with amorphous metal-oxide semiconductor (AOS) thin-?lm devices. In this study, we have developed a Hop?eld neural network with double-layer AOS thin-?lm devices as crosspoint-type synapse elements. Here, we propose modi?ed Hebbian learning done locally without extra control circuits, where the conductance deterioration of the crosspoint-type synapse elements can be employed as synaptic plasticity. In order to validate the fundamental operation of the neuromorphic system, ?rst, double-layer AOS thin-?lm devices as crosspoint-type synapse elements are actually fabricated, and it is found that the electric current continuously decreases along the bias time. Next, a Hop?eld neural network is really assembled using a ?eld-programmable gate array (FPGA) chip and the double-layer AOS thin-?lm devices, and it is con?rmed that a necessary function of the letter recognition is obtained after learning process. Once the fundamental operations are con?rmed, more advanced functions will be obtained by scaling up the devices and circuits. Therefore, it is expected the neuromorphic systems can be three-dimensional (3D) large-scale integration (LSI) chip, the machine size can be compact, power consumption can be low, and various functions of human brains will be obtained. What has been developed in this study will be the sole solution to realize them.
关键词： Neural network,Hop?eld neural network,Letter recognition,Arti?cial intelligence,Crosspoint-type synapse elements,Double-layer amorphous metal-oxide semiconductor (AOS) thin-?lm device,Modi?ed hebbian learning
更新于2025-09-09 09:28:46
-
[IEEE IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Valencia, Spain (2018.7.22-2018.7.27)] IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Deconv R-CNN for Small Object Detection on Remote Sensing Images
摘要： Small object detection has drawn increasing interest in computer vision and remote sensing image processing. The Region Proposal Network (RPN) methods (e.g., Faster R-CNN) have obtained promising detection accuracy with several hundred proposals. However, due to the pooling layers in the network structure of the deep model, precise localization of small-size object is still a hard problem. In this paper, we design a network with a deconvolution layer after the last convolution layer of base network for small target detection. We call our model DeconvR-CNN. In the experiment on a remote sensing image dataset, DeconvR-CNN reaches a much higher mean average precision (mAP) than Faster R-CNN.
关键词： Object detection,Small object,Convolutional neural network,R-CNN,Deconvolution
更新于2025-09-09 09:28:46
-
[IEEE IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Valencia, Spain (2018.7.22-2018.7.27)] IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Target detection in remote sensing image based on saliency computation of spiking neural network
摘要： Target detection is a-priori conditions for target tracking, classification, recognition, and scene understanding in Remote Sensing Image (RSI) analysis. However, the many traditional algorithms for target detection cannot perform well when the image resolution, especially for high-resolution RSIs, is change. Therefore, in this paper, we introduce a novel target detection algorithm based on the visual saliency of Spiking Neural Networks (SNN), which can efficiently detect the discriminative information from high-resolution RSIs to find targets by a saliency computing. As a result of this, it can provide an efficient and fast calculation method. The proposed visual saliency algorithm was applied to extensive experiments to detect the ship, and experimental results showed the outstanding performance for target detection on the optical RSI and synthetic aperture image.
关键词： visual saliency,target detection,spiking neural network,high-resolution RSI,saliency computation
更新于2025-09-09 09:28:46
-
[IEEE IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Valencia, Spain (2018.7.22-2018.7.27)] IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - CNN-Based Target Detection in Hyperspectral Imagery
摘要： This paper proposes a hyperspectral target detection framework with convolutional neural network (CNN). The number of training samples is first sufficiently enlarged by subtraction method to maximize the advantages of the multilayer CNN. Next, the CNN is given a target detection function by labelling the new pixels subtracted between target and background classes as 1, and the pixels subtracted between pixels within both the same and different background classes as 0. Finally, for each testing pixel, the difference between the central pixel and its adjacent pixels is input into the framework. If the testing pixel belongs to the target, the output score is close to the target label. Aircrafts and vehicles are selected as targets of interest in the experiment conducted to validate the proposed method. The experiment results show that the proposed method has an advantage over classic hyperspectral target detection algorithms in terms of precision and robustness.
关键词： Deep Learning,Convolutional Neural Network,Target Detection,Remote Sensing,Hyperspectral
更新于2025-09-09 09:28:46
-
[IEEE 2018 2nd IEEE Advanced Information Management,Communicates, Electronic and Automation Control Conference (IMCEC) - Xi'an (2018.5.25-2018.5.27)] 2018 2nd IEEE Advanced Information Management,Communicates,Electronic and Automation Control Conference (IMCEC) - Hyperspectral Target Detection with CNN Using Subtraction Model
摘要： Recently, the convolutional neural network (CNN) has been widely used in the fields of hyperspectral image (HSI) processing. In this paper, a CNN-based hyperspectral target detection framework is presented. And subtraction model is used to sufficiently enlarge the number of training samples. The subtraction model is built from twenty-eight manually selected objects in several AVIRIS date following three aspects: 1) The new pixel made by subtraction of any two pixels between 27 different classes is labelled as 0; 2) the new pixel made by subtraction of any two pixels within per class is labelled as 0; 3) the new pixel made by subtraction of any two pixels, in which one pixel is from the target class and the other is from background classes, is labelled as 1. Theoretically, if the pixel under test belongs to the target class, the output label of the CNN will be the same as the label of the target class. The experiment results on three images all indicate that the proposed CNN-based detector outperforms the classical hyperspectral target detection algorithms.
关键词： target detection,convolutional neural network,deep learning,hyperspectral imagery
更新于2025-09-09 09:28:46
-
[IEEE 2018 International Conference on 3D Vision (3DV) - Verona (2018.9.5-2018.9.8)] 2018 International Conference on 3D Vision (3DV) - MVDepthNet: Real-Time Multiview Depth Estimation Neural Network
摘要： Although deep neural networks have been widely applied to computer vision problems, extending them into multiview depth estimation is non-trivial. In this paper, we present MVDepthNet, a convolutional network to solve the depth estimation problem given several image-pose pairs from a localized monocular camera in neighbor viewpoints. Multiview observations are encoded in a cost volume and then combined with the reference image to estimate the depth map using an encoder-decoder network. By encoding the information from multiview observations into the cost volume, our method achieves real-time performance and the flexibility of traditional methods that can be applied regardless of the camera intrinsic parameters and the number of images. Geometric data augmentation is used to train MVDepthNet. We further apply MVDepthNet in a monocular dense mapping system that continuously estimates depth maps using a single localized moving camera. Experiments show that our method can generate depth maps efficiently and precisely.
关键词： convolutional neural network,cost volume,monocular dense mapping,multiview depth estimation,geometric data augmentation
更新于2025-09-09 09:28:46
-
[IEEE 2018 Condition Monitoring and Diagnosis (CMD) - Perth, WA (2018.9.23-2018.9.26)] 2018 Condition Monitoring and Diagnosis (CMD) - Pattern Recognition of Partial Discharge Image Based on One-dimensional Convolutional Neural Network
摘要： Big data platforms and centers are ubiquitous today where a large amount of unstructured data on site such as is accumulated. For structured data, partial discharge pattern recognition method has been extensively studied, whereas traditional methods can not be directly applied to unstructured data. To this end, a time-domain waveform pattern recognition method based on one- dimensional convolutional neural network (CNN) is proposed. Image processing techniques are applied to obtain one- dimensional characteristics of the waveform. Based on deep learning, the network is constructed for pattern recognition straight forwardly. Through on site detection and simulation experiments, image data sets of five partial discharge defects are established and comparative experiments are conducted. Experimental results show that the proposed method can successfully perform pattern recognition with applications in work of data mining and data utilization. Under the same complexity, it is also with higher accuracy comparing to two- dimensional CNN. Furthermore, the method autonomously extrapolates features without manual extraction, which achieves low experimental complexity and robustness simultaneously.
关键词： pattern recognition,image,partial discharge,convolutional neural network(CNN)
更新于2025-09-09 09:28:46
-
[IEEE 2018 25th IEEE International Conference on Image Processing (ICIP) - Athens, Greece (2018.10.7-2018.10.10)] 2018 25th IEEE International Conference on Image Processing (ICIP) - Only-Reference Video Quality Assessment for Video Coding Using Convolutional Neural Network
摘要： Conventional video quality assessment methods are either full-, reduced-, or no-reference methods that need to access decoded videos. Hence, to calculate quality of decoded video in video coding regarding an image/video quality metric, complete encoding and decoding have to executed, which is computationally expensive. To address this problem, we propose to estimate quality of decoded videos from the original video only (i.e., only-reference) using convolutional neural network, as if the original video is encoded using a range of quantization parameter. The proposed network is shallow and can be trained to estimate various video quality metrics. Furthermore, among potential rate control applications using the proposed network, we demonstrate achieving a targeted decoded-video quality by selecting a proper quantization parameter before actually encoding.
关键词： only-reference,Video quality assessment,convolutional neural network,video coding
更新于2025-09-09 09:28:46