The evaluation and manipulation of robot motion are essential for effective motion planning. Similar to audio and video data, motion data requires transform operations and a merit function when utilizing supervised an...
详细信息
bigdata wave has led to a rapid increase in the amount of data being collected by organizations. While the accuracy and reliability of prediction models are often prioritized, the quality of the collected data is fre...
详细信息
bigdata wave has led to a rapid increase in the amount of data being collected by organizations. While the accuracy and reliability of prediction models are often prioritized, the quality of the collected data is frequently overlooked. Poor data quality can result in the common problem of ‘garbage in, garbage out’. Traditional measures of data quality, such as accuracy, consistency, completeness, and timeliness, are no longer adequate in the era of bigdata. Therefore, this paper proposes a taxonomy of data quality dimensions specifically for bigdata, addressing emerging challenges by formulating 20 dimensions and categorizing them into four distinct categories.
Deep hashing is an appealing approach for large-scale image retrieval. Most existing supervised deep hashing methods learn hash functions using pairwise or triple image similarities in randomly sampled mini-batches. T...
Deep hashing is an appealing approach for large-scale image retrieval. Most existing supervised deep hashing methods learn hash functions using pairwise or triple image similarities in randomly sampled mini-batches. They suffer from low training efficiency, insufficient coverage of data distribution, and pair imbalance problems. Recently, central similarity quantization (CSQ) attacks the above problems by using “hash centers” as a global similarity metric, which encourages the hash codes of similar images to approach their common hash center and distance themselves from other hash centers. Although achieving SOTA retrieval performance, CSQ falls short of a worst-case guarantee on the minimal distance between its constructed hash centers, i.e. the hash centers can be arbitrarily close. This paper presents an optimization method that finds hash centers with a constraint on the minimal distance between any pair of hash centers, which is non-trivial due to the non-convex nature of the problem. More importantly, we adopt the Gilbert-Varshamov bound from coding theory, which helps us to obtain a large minimal distance while ensuring the empirical feasibility of our optimization approach. With these clearly-separated hash centers, each is assigned to one image class, we propose several effective loss functions to train deep hashing networks. Extensive experiments on three datasets for image retrieval demonstrate that the proposed method achieves superior retrieval performance over the state-of-the-art deep hashing methods.
— A novel planar balanced-to-balanced (BTB) microstrip filtering crossover with high isolation and common-mode (CM) suppression is proposed in this paper. The circuit structure is based on microstrip transmission lin...
详细信息
In this work, we propose a novel method for text-to-image generation, combining the techniques of stable diffusion and large language model. Our approach improves prompt representation using large language model, with...
详细信息
ISBN:
(数字)9798350375107
ISBN:
(纸本)9798350375114
In this work, we propose a novel method for text-to-image generation, combining the techniques of stable diffusion and large language model. Our approach improves prompt representation using large language model, with key tokens based on defects in the observed images. These key tokens draws inspiration from prompt learning. We improve synthesized images by using better prompts with key tokens from a large language model. Experiments show that ourapproach achieves better performance compared to existing methods.
This paper presents an efficient and scalable incomplete multi-view clustering method, referred to as Enhanced Dictionary-Induced tenSorized incomplete multi-view clustering with Gaussian errOr raNk minimization (EDIS...
This paper presents an efficient and scalable incomplete multi-view clustering method, referred to as Enhanced Dictionary-Induced tenSorized incomplete multi-view clustering with Gaussian errOr raNk minimization (EDISON). Specifically, EDISON employs an enhanced dictionary representation strategy as the foundation for inferring missing data and constructing anchor graphs, ensuring robustness to less-than-ideal data and maintaining high computational efficiency. Additionally, we introduce Gaussian error rank as a concise approximation of the true tensor rank, facilitating a comprehensive exploration of the diverse information encapsulated by various singular values in tensor data. Furthermore, we integrate a hyper-anchor graph Laplacian manifold regularization into the tensor representation, allowing for the simultaneous utilization of inter-view high-order correlations and intra-view local correlations. Extensive experiments demonstrate the superiority of the EDISON model in both effectiveness and efficiency compared to SOTA methods.
In the process of drawing architectural drawings with AutoCAD software, enterprises will produce a large number of CAD drawings in DWG format. The tables of these CAD drawings contain rich textual information. These t...
详细信息
ISBN:
(数字)9798350368604
ISBN:
(纸本)9798350368611
In the process of drawing architectural drawings with AutoCAD software, enterprises will produce a large number of CAD drawings in DWG format. The tables of these CAD drawings contain rich textual information. These text information constitute the key data base of enterprise systematic management drawings. At present, the systematic management of drawing data needs to collect these sheet data, but the sheet data collection mainly depends on manual input, which leads to high error rate of data collection, huge workload, long compilation cycle and low work efficiency. The purpose of this paper is to discuss how to deal with these tables efficiently. Firstly, based on Revit secondary development, the frame recognition algorithm is proposed, which confirms the inner and outer frame by finding the largest rectangle and the second largest rectangle. In addition, the Teigha class library is used to extract text, so as to extract key data, so as to promote the collection, reuse and systematic management of product data. Through the processing of drawing frames and forms, it lays a good foundation for subsequent component identification. The method proposed in this paper has strong versatility and adaptability, and can basically realize the recognition of CAD drawing frame and form, and extract form text, and achieve higher accuracy and efficiency.
Colorectal cancer is a common malignancy. In colonoscopy images, computer-assisted polyp segmentation helps doctors diagnose and treat disorders more precisely. In recent years, some methods based on deep convolutiona...
详细信息
In this modern world due to Road traffic, many people are unable to reach their destination at the correct time. For example, if a person needed to reach the hospital in critical condition due to road traffic, they ar...
详细信息
Knowledge bases (KBs), which store high-quality information, are crucial for many applications, such as enhancing search results and serving as external sources for data cleaning. Not surprisingly, there exist outdate...
详细信息
暂无评论