Recently, some studies have shown that semantic and distortion representations both benefit the evaluation of image quality. However, the images of existing synthetic distortion databases are annotated with subjective...
详细信息
Recently, some studies have shown that semantic and distortion representations both benefit the evaluation of image quality. However, the images of existing synthetic distortion databases are annotated with subjective quality scores and distortion types, lacking labels with semantic objects. Therefore, it is virtually infeasible to learn the representations of image semantics and distortion by co-guiding with semantic and distortion labels. To address this issue, we propose a dual-perception network (DPNet) via an end-to-end multi-task learning method, where knowledge distillation is lever-aged as a semantic label-free strategy. Specifically, semantic representation derived from pre-trained ResNet152 is applied to supervise the output of DPNet, while the output is utilized to construct a distortion recognition task. In this way, image semantics and distortion can be hybridly represented in an identical feature map. Finally, image quality is regressed based on the hybrid representations. Experimental results conducted on five benchmark databases validate that the proposed method can achieve state-of-the-art performance.
Type II Solar Radio Bursts (SRBs) are the result of particle acceleration by shock waves in the solar corona and interplanetary medium. The shocks are created by solar eruptions involving coronal mass ejections travel...
Type II Solar Radio Bursts (SRBs) are the result of particle acceleration by shock waves in the solar corona and interplanetary medium. The shocks are created by solar eruptions involving coronal mass ejections traveling at super-alfvenic speeds. The automatic detection, classification, and segmentation of such radio bursts is a challenge in solar radio physics due to their heterogeneous form. Large data rates produced by cutting-edge radio telescopes like the LOw-Frequency ARray (LOFAR) have made SRB detection and classification more feasible in recent years. In this study, we use a Generative Adversarial Network (GAN) to simulate Type II SRBs and then use this simulated data as a training set for a Mask R-CNN to detect and segment Type II SRBs automatically. Using this multi-model approach, we can accurately detect and segment Type II SRBs with a mean Average Precision (mAP) score of 78.90%.
Universal Information Extraction (UIE) has been introduced as a unified framework for various Information Extraction (IE) tasks and has achieved widespread success. Despite this, UIE models have limitations. For examp...
详细信息
Reasoning about strategic abilities is key to AI systems comprising multiple agents, which provide a unified framework for formalizing various problems in game theory, social choice theory, etc. In this work, we propo...
详细信息
In the era of Big Data, Information can be generated, extracted, and utilized in diverse ways. In business, information about business capabilities can be a crucial aspect in understanding the strengths and competenci...
In the era of Big Data, Information can be generated, extracted, and utilized in diverse ways. In business, information about business capabilities can be a crucial aspect in understanding the strengths and competencies of organizations. This Information can empower stakeholders to make informed decisions regarding partnerships, acquisitions, and market positioning. Additionally, by comparing capabilities across companies and industries, organizations can identify their competitive advantages, industry trends, potential disruptions, and effective resource allocations, leading to improved strategies and better market opportunities. Traditional methods for extracting business capabilities primarily rely on keyword matching or rule-based techniques, which often fail to capture the complex relationships between words in a sentence. This paper presents a comprehensive study on extracting business capabilities using dependency parsing techniques in natural language processing (NLP) to overcome limitations of the traditional methods and improve the accuracy of business capability extraction. Dependency parsing utilizes pre-trained language models to analyze the grammatical structure of sentences, identifying the dependencies between words and capturing their relationships. The feasibility and effectiveness of the proposed approach is demonstrated through experiments conducted on more than 40K real-world company descriptions crawled from Wikipedia.
Ultra high-speed and reliable next-generation 6G mobile networks are recognized as key enablers for many innovative scenarios in smart cities - from vehicular use cases and surveillance to healthcare. However, deploym...
详细信息
Federated Learning (FL) is a framework that aims to learn in a distributed and privacy preserving manner. It often works with a central server coordinating multiple clients to jointly train a global model, without the...
详细信息
Conventional active learning approaches for hyperspectral image classification (HSIC) have limitations such as incrementally growing training sets without considering class structure and heterogeneity within existing ...
详细信息
ISBN:
(数字)9798350349399
ISBN:
(纸本)9798350349405
Conventional active learning approaches for hyperspectral image classification (HSIC) have limitations such as incrementally growing training sets without considering class structure and heterogeneity within existing and new samples. Additionally, there is limited research leveraging both spectral and spatial information jointly, and stopping criteria are not well established. This study presents a novel fuzzybased spatial-spectral Within and Between method (FLG) for preserving local and global class discriminative information. The method first explores spatial fuzziness to identify misclassified samples. It then computes total within-class and between-class information locally and globally. This information is integrated into a discriminative objective function to selectively query heterogeneous samples, mitigating randomness among training data. Experimental results on benchmark Hyperspectral datasets demonstrate the FLG improves classification accuracy across generative, extreme learning machine, and sparse multinomial logistic regression models by jointly exploiting spectral and spatial information to expand labeled training sets strategically.
Sequential recommendations have drawn significant attention in modeling the user’s historical behaviors to predict the next item. With the booming development of multimodal data (e.g., image, text) on internet platfo...
详细信息
暂无评论