Under snowy weather conditions, cameras are prone to the interference of snow and can severely reduce the quality of the captured images, which will affect the computervision performance greatly. Since no temporal in...
详细信息
ISBN:
(纸本)9781510666313;9781510666320
Under snowy weather conditions, cameras are prone to the interference of snow and can severely reduce the quality of the captured images, which will affect the computervision performance greatly. Since no temporal information can be exploited, snow removal from single image is a challenging problem. In this paper, a novel snow removal method from single image was proposed by designing a kind of multi-scale imageprocessing framework both in the spatial and frequency domain. Firstly, the input snowy image was decomposed into detailed sub-images and approximate parts by the Laplacian pyramid transform. Secondly, the approximate part is decomposed again into the background and detailed sub-image by the edge-preserving and structure-preserving image smoothing filter. After that, the non-subsampled shearlet transform was introduced to detect snowflakes within the frequency domain of the detailed sub-images, while mathematical morphological filtering was adopted to remove the labeled snowflakes within their spatial domain. Finally, the desnowing image was obtained by the inverse Laplacian pyramid transform. Experiments on real-world snowy images show that the proposed method produces better results than those of other state-of-the-art methods.
Classification algorithm based on deep learning is the main technology for computer-aided intelligent diagnosis of breast pathological images. The existing deep learning algorithms rarely pay attention to multi-scale ...
详细信息
Saving algorithmic overhead for realistic effect generation is always a hot topic in graphics research. The algorithm in this paper applies the method of the Generative Adversarial Network (GAN) to the study of realis...
详细信息
ISBN:
(纸本)9781510666313;9781510666320
Saving algorithmic overhead for realistic effect generation is always a hot topic in graphics research. The algorithm in this paper applies the method of the Generative Adversarial Network (GAN) to the study of realistic computergraphics, proposing a new solution to reduce the algorithm overhead. We adopt conditional GAN and add an additional module before the generator to provide the lighting information of the current scene. This extra module represents the illuminating direction with grayscale gradient maps of different angles helping the network get lighting information through image input. In addition, we optimize the loss function by adding the L1 loss and the feature perception loss to improve the generative effect of our network. In the feature perception loss, we use a pre-trained VGG network to calculate the detail feature gap between images, to help the model generate images with better light and shadow effects. Our algorithm can add realistic effects to the existing coarse-rendered image according to the lighting conditions and obtain the corresponding fine-rendered image in single light scenes. The extensive experimental results show that our algorithm has a good post-processing effect on realistic rendering, and the time overhead of the algorithm is independent of the complexity of the scene. From the robustness test, we can know that our network also has a good generalization ability.
The proceedings contain 17 papers. The special focus in this conference is on computervision, graphics, and imageprocessing. The topics include: The Ikshana Hypothesis of Human Scene Understanding;worst-Case Adversa...
ISBN:
(纸本)9789811941351
The proceedings contain 17 papers. The special focus in this conference is on computervision, graphics, and imageprocessing. The topics include: The Ikshana Hypothesis of Human Scene Understanding;worst-Case Adversarial Perturbation and Effect of Feature Normalization on Max-Margin Multi-label Classifiers;Catch Me if You Can: A Novel Task for Detection of Covert Geo-Locations (CGL);MATIC: Memory-Guided Adaptive Transformer for image Captioning;Semantic Map Injected GAN Training for image-to-image Translation;textGen3D: A Real-Time 3D-Mesh Generation with Intersecting Contours for Text;comparative Analysis of Neural Architecture Search Methods for Classification of Cultural Heritage Sites;heritage Representation of Kashi Vishweshwar Temple at Kalabgoor, Telangana with Augmented Reality Application Using Photogrammetry;augmented Data as an Auxiliary Plug-In Toward Categorization of Crowdsourced Heritage Data;evolution of Bagbazar Street Through Visibility Graph Analysis (1746–2020);mapping Archaeological Remains of 14th Century Fort of Jahanpanah Using Geospatial Analysis;spatial Analysis and 3d Mapping Historic Landscapes—Implications of Adopting an Integrated Approach in Simulation and Visualization of Landscapes;HSADML: Hyper-Sphere Angular Deep Metric Based Learning for Brain Tumor Classification;model Compression Based Lightweight Online Signature Verification Framework.
Earth observation satellites provide us with ample amount of raw data for land cover analysis. However, annotating these data is a cumbersome process, subjected to human error which compel us to shift from supervised ...
详细信息
ISBN:
(纸本)9798400716256
Earth observation satellites provide us with ample amount of raw data for land cover analysis. However, annotating these data is a cumbersome process, subjected to human error which compel us to shift from supervised to unsupervised techniques. Although clustering methods are being widely used for the past few years in the field of remote sensing, identifying the number of fine-grain classes present in a region remain a challenging problem. Therefore, we propose a rule-based and neural-network learning technique that can divide the pixels into three standard classes, water bodies, vegetation and vegetation-void. These classes are easier to identify and is not region-specific. Later we apply fine-grain clustering on each of these classes to segregate them into finer groups. Our clustering method identifies the appropriate number of fine-grain classes present in a specific region.
The important features that enable computervision in autonomous vehicle technology and infotainment function are imageprocessing and object identification. image segmentation is the preliminary step of any image pro...
详细信息
The visual quality of nighttime photographs diminishes greatly due to low contrast and high noise. We need a robust image enhancement methodology to improve such low-light images close to standard daylight images. Due...
详细信息
Currency fraud now possess a serious risk to both people's livelihoods and the nation's economy. Due to improvements, in printing and scanning technology, currency duplication is still a growing worry for gove...
详细信息
The proceedings contain 140 papers. The special focus in this conference is on computervision and imageprocessing. The topics include: A Novel Framework for Cognitive Load Estimation from Electroencephalog...
ISBN:
(纸本)9783031581809
The proceedings contain 140 papers. The special focus in this conference is on computervision and imageprocessing. The topics include: A Novel Framework for Cognitive Load Estimation from Electroencephalogram Signals Utilizing Sparse Representation of Brain Connectivity;MCCNet: A Multi-scale Cross Connection Network for Low-Light image Enhancement;feature Fusion and Multi-head Attention Based Hindi Captioner;Payload Length and Location Identification Using a Novel CNN and Re-embedding Strategy in Stego images Created by Content Adaptive Steganographic Algorithms;ConvMTL: Multi-task Learning via Self-supervised Learning for Simultaneous Dense Predictions;driver Drowsiness Detection Using vision Transformer;automated Cricket Commentary Generation for Videos;image Captioning with Visual Positional Embedding and Bi-linear Pooling;Cartilage Segmentation from MRI images Towards Prediction of Osteoarthritis;Reinforcement Algorithm-Guided ROI Extraction of Fingerprint Biometric Data;improved Metric Space for Shape Correspondence;an Integrated Approach: Combining GrabCut and Contour-Matching for Hand Gesture Segmentation in indian Sign Language;identity Preserved Expressive Talking Faces with Synchrony;isolated Sign Language Recognition Using Deep Learning;a Gradient-Based Approach to Interpreting Visual Encoding Models;adjust Your Focus: Defocus Deblurring from Dual-Pixel images Using Explicit Multi-Scale Cross-Correlation;Efficient Seizure Prediction from images of EEG Signals Using Convolutional Neural Network;coping with Increased Levels of Label Noise in Facial Expression Recognition;experimental Evaluation of Needle Tip Prediction Using Kalman Filtering Approach;a Deep Face Antispoofing System with Hardware Implementation for Real-Time Applications;Bit Plane Segmentation and LBP-Based Coverless Video Steganography for Secure Data Transmission;a Deep Learning Approach to Enhance Semantic Segmentation of Bacteria and Pus Cells from Microscopic Urine Smear images Using
The proceedings contain 140 papers. The special focus in this conference is on computervision and imageprocessing. The topics include: A Novel Framework for Cognitive Load Estimation from Electroencephalog...
ISBN:
(纸本)9783031581731
The proceedings contain 140 papers. The special focus in this conference is on computervision and imageprocessing. The topics include: A Novel Framework for Cognitive Load Estimation from Electroencephalogram Signals Utilizing Sparse Representation of Brain Connectivity;MCCNet: A Multi-scale Cross Connection Network for Low-Light image Enhancement;feature Fusion and Multi-head Attention Based Hindi Captioner;Payload Length and Location Identification Using a Novel CNN and Re-embedding Strategy in Stego images Created by Content Adaptive Steganographic Algorithms;ConvMTL: Multi-task Learning via Self-supervised Learning for Simultaneous Dense Predictions;driver Drowsiness Detection Using vision Transformer;automated Cricket Commentary Generation for Videos;image Captioning with Visual Positional Embedding and Bi-linear Pooling;Cartilage Segmentation from MRI images Towards Prediction of Osteoarthritis;Reinforcement Algorithm-Guided ROI Extraction of Fingerprint Biometric Data;improved Metric Space for Shape Correspondence;an Integrated Approach: Combining GrabCut and Contour-Matching for Hand Gesture Segmentation in indian Sign Language;identity Preserved Expressive Talking Faces with Synchrony;isolated Sign Language Recognition Using Deep Learning;a Gradient-Based Approach to Interpreting Visual Encoding Models;adjust Your Focus: Defocus Deblurring from Dual-Pixel images Using Explicit Multi-Scale Cross-Correlation;Efficient Seizure Prediction from images of EEG Signals Using Convolutional Neural Network;coping with Increased Levels of Label Noise in Facial Expression Recognition;experimental Evaluation of Needle Tip Prediction Using Kalman Filtering Approach;a Deep Face Antispoofing System with Hardware Implementation for Real-Time Applications;Bit Plane Segmentation and LBP-Based Coverless Video Steganography for Secure Data Transmission;a Deep Learning Approach to Enhance Semantic Segmentation of Bacteria and Pus Cells from Microscopic Urine Smear images Using
暂无评论