- 标题
- 摘要
- 关键词
- 实验方案
- 产品
-
The Spectral-Spatial Joint Learning for Change Detection in Multispectral Imagery
摘要： Change detection is one of the most important applications in the remote sensing domain. More and more attention is focused on deep neural network based change detection methods. However, many deep neural networks based methods did not take both the spectral and spatial information into account. Moreover, the underlying information of fused features is not fully explored. To address the above-mentioned problems, a Spectral-Spatial Joint Learning Network (SSJLN) is proposed. SSJLN contains three parts: spectral-spatial joint representation, feature fusion, and discrimination learning. First, the spectral-spatial joint representation is extracted from the network similar to the Siamese CNN (S-CNN). Second, the above-extracted features are fused to represent the difference information that proves to be effective for the change detection task. Third, the discrimination learning is presented to explore the underlying information of obtained fused features to better represent the discrimination. Moreover, we present a new loss function that considers both the losses of the spectral-spatial joint representation procedure and the discrimination learning procedure. The effectiveness of our proposed SSJLN is verified on four real data sets. Extensive experimental results show that our proposed SSJLN can outperform the other state-of-the-art change detection methods.
关键词： discrimination learning,feature fusion,change detection,spectral-spatial representation,multispectral imagery,Siamese CNN
更新于2025-09-19 17:15:36
-
A multi-faceted CNN architecture for automatic classification of mobile LiDAR data and an algorithm to reproduce point cloud samples for enhanced training
摘要： Mobile Laser Scanning (MLS) data of outdoor environment are typically characterised by occlusion, noise, clutter, large data size and high quantum of information which makes their classification a challenging problem. This paper presents three deep Convolutional Neural Network (CNN) architectures in three dimension (3D), namely single CNN (SCN), multi-faceted CNN (MFC) and MFC with reproduction (MFCR) for automatic classification of MLS data. The MFC uses multiple facets of an MLS sample as inputs to different SCNs, thus providing additional information during classification. The MFC, once trained, is used to reproduce additional samples with the help of existing samples. The reproduced samples are employed to further refine the MFC training parameters, thus giving a new method called MFCR. The three architectures are evaluated on an ensemble of 3D outdoor MLS data consisting of four classes, i.e. tree, pole, house and ground covered with low vegetation along with car samples from KITTI dataset. The total accuracy and kappa values of classifications reached up to (i) 86.0% and 81.3% for the SCN (ii) 94.3% and 92.4% for the MFC and (iii) 96.0% and 94.6% for the MFCR, respectively. The paper has demonstrated the use of multiple facets to significantly improve classification accuracy over the SCN. Finally, a unique approach has been developed for reproduction of samples which has shown potential to improve the accuracy of classification. Unlike previous works on the use of CNN for structured point cloud of indoor objects, this work shows the utility of different proposed CNN architectures for classification of varieties of outdoor objects, viz., tree, pole, house and ground which are captured as unstructured point cloud by MLS.
关键词： Sample reproduction,Mobile Laser Scanning (MLS),Automatic classification,Convolutional Neural Network (CNN)
更新于2025-09-19 17:15:36
-
[IEEE 2018 7th European Workshop on Visual Information Processing (EUVIP) - Tampere, Finland (2018.11.26-2018.11.28)] 2018 7th European Workshop on Visual Information Processing (EUVIP) - A Hybrid Approach to Hand Detection and Type Classification in Upper-Body Videos
摘要： Detection of hands in videos and their classification into left and right types are crucial in various human-computer interaction and data mining systems. A variety of effective deep learning methods have been proposed for this task, such as region-based convolutional neural networks (R-CNNs), however the large number of their proposal windows per frame deem them computationally intensive. For this purpose we propose a hybrid approach that is based on substituting the 'selective search' R-CNN module by an image processing pipeline assuming visibility of the facial region, as for example in signing and cued speech videos. Our system comprises two main phases: preprocessing and classification. In the preprocessing stage we incorporate facial information, obtained by an AdaBoost face detector, into a skin-tone based segmentation scheme that drives Kalman filtering based hand tracking, generating very few candidate windows. During classification, the extracted proposal regions are fed to a CNN for hand detection and type classification. Evaluation of the proposed hybrid approach on four well-known datasets of gestures and signing demonstrates its superior accuracy and computational efficiency over the R-CNN and its variants.
关键词： region-based convolutional neural network (R-CNN),hand type classification,Hand detection,AdaBoost face detection,Kalman filtering
更新于2025-09-19 17:15:36
-
[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) - Bi-Directional Vectors from Apex in CNN for Micro-Expression Recognition
摘要： The impressive performance of utilizing deep learning or neural network has attracted much attention in both the industry and research communities, especially towards computer vision aspect related applications. Despite its superior capability of learning, generalization and interpretation on various form of input, micro-expression analysis field is yet remains new in applying this kind of computing system in automated expression recognition system. A new feature extractor, BiVACNN is presented in this paper, where it first estimates the optical flow fields from the apex frame, then encode the flow fields features using CNN. Concretely, the proposed method consists of three stages: apex frame acquisition, multivariate features formation and feature learning using CNN. In the multivariate features formation stage, we attempt to derive six distinct features from the apex details, which include: the apex itself, difference between the apex and onset frames, horizontal optical flow, vertical optical flow, magnitude and orientation. It is demonstrated that utilizing the horizontal and vertical optical flow capable to achieve 80% recognition accuracy in CASME II and SMIC-HS databases.
关键词： CNN,micro-expression,apex,optical flow,recognition
更新于2025-09-19 17:15:36
-
[IEEE IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Valencia (2018.7.22-2018.7.27)] IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Shipnet for Semantic Segmentation on VHR Maritime Imagery
摘要： For VHR maritime images, sematic segmentation is a new research hotspot and plays an important role in coast line navigation, resource management and territory protection. Without enough labeled training data, it is a challenge to separate small objects on a large scale while segment the big area clearly. To deal with it, we propose a novel ShipNet and design a weighted loss function for simultaneous sea-land segmentation and ship detection. To prove the proposed method, we also built and opened a new dataset to the community which contains VHR multiscale maritime images. Compared with the FCN and ResNet, the proposed method got much better F1 scores 85.90% for ship class and 97.54% overall accuracy. Compared with multiscale FCN, the ShipNet could obtain details results like sharp edges. Even for images with bad quality, the ShipNet could also keep robust and get good results.
关键词： CNN,ship detection,Sea-land segmentation,remote sensing image
更新于2025-09-19 17:15:36
-
[IEEE 2018 21st International Conference on Intelligent Transportation Systems (ITSC) - Maui, HI (2018.11.4-2018.11.7)] 2018 21st International Conference on Intelligent Transportation Systems (ITSC) - Visual Quality Enhancement Of Images Under Adverse Weather Conditions
摘要： The visual quality of an image captured by vision systems can degrade significantly under adverse weather conditions. In this paper we propose a deep learning based solution to improve the visual quality of images captured under rainy and foggy circumstances, which are among the prominent and common weather conditions that attribute to bad image quality. Our convolutional neural network (CNN), NVDeHazeNet learns to predict both the original signal as well as the atmospheric light to finally restore image quality. It outperforms the existing state of the art methods by evaluation on both synthetic data as well as real world hazy images. The deraining CNN, NVDeRainNet shows similar performance on existing rain datasets as the state of the art. On natural rain images NVDeRainNet shows better than state of the art performance. We show the use of perceptual loss to improve the visual quality of results. These networks require considerable amount of data under adverse weather conditions and their respective ground truth for training. For this purpose we use a weather simulation framework to simulate synthetic rainy and foggy environments. This data is augmented with existing rain datasets to train the networks.
关键词： deraining,dehazing,cnn,deep learning
更新于2025-09-19 17:15:36
-
[IEEE 2018 26th Telecommunications Forum (TELFOR) - Belgrade, Serbia (2018.11.20-2018.11.21)] 2018 26th Telecommunications Forum (TELFOR) - RGB-NIR Demosaicing Using Deep Residual U-Net
摘要： Multi-spectral image acquisition brings numerous potential benefits in computer vision and image processing applications. Single-sensor acquisition helps to overcome problems with misalignments occurring in multiple-sensor acquisition. However, the single-sensor approach poses the problem of interpolation of missing values. In this paper we propose an adapted version of a residual U-Net, with application in demosaicing. The experiments show that the proposed method achieves state-of-the-art results, and has good generalization capabilities to different color filter array patterns.
关键词： single-sensor,CNN,RGB-NIR,CFA,deep learning,demosaicing,U-Net
更新于2025-09-19 17:15:36
-
Effect of denoising on hyperspectral image classification using deep networks and kernel methods
摘要： Hyperspectral Image (HSI) store the re?ectance values of a single scene or object in several continuous bands of electromagnetic spectrum. When the image is recorded, the information in some of the spectral bands gets mixed with noise. The classi?cation accuracy of hyperspectral image varies inversely with the quantity and nature of noise present in the cluster of spectral bands. Thus, denoising is a fundamental prerequisite in image processing applications like classi?cation, unmixing, etc. In this paper, we compare the effect of denoising via classi?cation using Vectorized Convolutional Neural Network (VCNN), kernel based Support Vector Machine (SVM) and Grand Uni?ed Regularized Least Squares (GURLS) classi?ers. The classi?ers are provided with raw data (without denoising) and denoised data using spectral and spatial Least Square (LS) techniques. The data given to the network are in the form of pixels, so we call the convolutional neural network (CNN) as VCNN. The experiments are performed on three standard HSI datasets. The performance of the classi?ers are evaluated based on overall and class-wise accuracy.
关键词： CNN,Least Square Denoising,IBBC,GURLS,LIBSVM,Hyperspectral Image
更新于2025-09-19 17:15:36
-
Deep Convolutional Neural Network for HEp-2 Fluorescence Intensity Classification
摘要： Indirect ImmunoFluorescence (IIF) assays are recommended as the gold standard method for detection of antinuclear antibodies (ANAs), which are of considerable importance in the diagnosis of autoimmune diseases. Fluorescence intensity analysis is very often complex, and depending on the capabilities of the operator, the association with incorrect classes is statistically easy. In this paper, we present a Convolutional Neural Network (CNN) system to classify positive/negative fluorescence intensity of HEp-2 IIF images, which is important for autoimmune diseases diagnosis. The method uses the best known pre-trained CNNs to extract features and a support vector machine (SVM) classifier for the final association to the positive or negative classes. This system has been developed and the classifier was trained on a database implemented by the AIDA (AutoImmunité, Diagnostic Assisté par ordinateur) project. The method proposed here has been tested on a public part of the same database, consisting of 2080 IIF images. The performance analysis showed an accuracy of fluorescent intensity around 93%. The results have been evaluated by comparing them with some of the most representative state-of-the-art works, demonstrating the quality of the system in the intensity classification of HEp-2 images.
关键词： autoimmune diseases,accuracy,SVM,receiver operating characteristic (ROC) curve,Convolutional Neural Network (CNN),IIF images
更新于2025-09-19 17:15:36
-
RAMS: Remote and automatic mammogram screening
摘要： About one in eight women in the U.S. will develop invasive breast cancer at some point in life. Breast cancer is the most common cancer found in women and if it is identified at an early stage by the use of mammograms, x-ray images of the breast, then the chances of successful treatment can be high. Typically, mammograms are screened by radiologists who determine whether a biopsy is necessary to ascertain the presence of cancer. Although historical screening methods have been effective, recent advances in computer vision and web technologies may be able to improve the accuracy, speed, cost, and accessibility of mammogram screenings. We propose a total screening solution comprised of three main components: a web service for uploading images and reviewing results, a machine learning algorithm for accepting or rejecting images as valid mammograms, and an artificial neural network for locating potential malignancies. Once an image is uploaded to our web service, an image acceptor determines whether or not the image is a mammogram. The image acceptor is primarily a one-class SVM built on features derived with a variational autoencoder. If an image is accepted as a mammogram, the malignancy identifier, a ResNet-101 Faster R-CNN, will locate tumors within the mammogram. On test data, the image acceptor had only 2 misclassifications out of 410 mammograms and 2 misclassifications out of 1,640 non-mammograms while the malignancy identifier achieved 0.951 AUROC when tested on BI-RADS 1, 5, and 6 images from the INbreast dataset.
关键词： Faster R-CNN,SVM,Deep Learning,DDSM,Convolutional,TensorFlow,INbreast,Mammograms,Telemedicine,Artificial Neural Network
更新于2025-09-19 17:15:36