Tire tracks are crucial evidence for investigating and identifying traffic accident scenes. Intelligent recognition of tire track images enables rapid and precise identification of suspicious vehicles, assisting inves...
详细信息
Tire tracks are crucial evidence for investigating and identifying traffic accident scenes. Intelligent recognition of tire track images enables rapid and precise identification of suspicious vehicles, assisting investigators in determining accident causes and liabilities. This enhances both the efficiency and accuracy of case resolution. However, tire track images are complex and diverse, significantly influenced by environmental factors such as lighting and road conditions. Moreover, the limited availability of tire track image samples poses a challenge for training deep learning models in practical applications. To address these challenges, particularly the limited generalization ability of few-shot metric learning models on unseen categories and their poor recognition rates in open-world environments, we propose a novel framework called TireNet for few-shot tire track recognition. First, we adopt a residual network integrated with coordinate attention as the backbone for feature extraction. Secondly, the context-aware features of the support set and the query set are extracted separately through attention-based bi-directional long short-term memory model. Subsequently, the cosine similarity between the features of the support set and the query set is calculated to determine the category of the query image. Finally, to address the class imbalance issue in the tire track image dataset, we utilize an improved Focal Loss for gradient updates. This enables the model to focus more on difficult samples during training, thereby enhancing the training efficiency, particularly in scenarios involving few-shot learning and open-world environments. To validate the proposed method, we constructed a tire track image dataset of 1700 samples, covering various environmental conditions, providing a rich resource for tire track recognition research. Experimental results demonstrate that the proposed method significantly outperforms few-shot image classification and classic machine
After an overview of visualizations to explore temporal patterns, we will focus on interfaces for discovering temporal event patterns in electronic health records. Specifying event sequence queries is challenging even...
详细信息
ISBN:
(纸本)9781568817125
After an overview of visualizations to explore temporal patterns, we will focus on interfaces for discovering temporal event patterns in electronic health records. Specifying event sequence queries is challenging even for skilled computer professionals familiar with SQL. Our novel interactive search strategies allow for aligning records on important events, ranking, and filtering combined with grouping of results to find common or rare events. A second approach is to use query-by-example, in which users specify a pattern and see a similarity-ranked list of results, but the similarity measure needs to be customized for different needs. Temporal summaries allow comparisons between groups. We will discuss the methods we use to evaluate the usefulness of our interfaces through collaborations with clinicians and hospital administrators on case studies. Finally, application of the techniques to other domains will be discussed.
The goal of this workshop is to acquire conceptual and practical skills in developing thin, flexible and customizable physical user interfaces with printed electronics for interactive devices and objects. The workshop...
详细信息
The goal of this workshop is to acquire conceptual and practical skills in developing thin, flexible and customizable physical user interfaces with printed electronics for interactive devices and objects. The workshop will cover personalized digital design of printed electronics, basics of different sensor types and actuators, and prototyping of printed electronics with conductive inkjet printing.
BottlePrint is a fabrication system that allows users to produce large-scale objects on desktop-scale fabrication ma-chines. The key idea behind bottlePrint is to complement 3D printing with ready-made objects, in our...
详细信息
Big Data analytics and Artificial Intelligence systems derive non-intuitive and often unverifiable inferences about individuals’ behaviors, preferences, and private lives. Drawing on diverse, feature-rich datasets of...
详细信息
Research indicates that less than 2% of the population reads license agreements during software installation [12]. To address this problem, we developed textured agreements, visually redesigned agreements that employ ...
详细信息
Networks have remained a challenge for information retrieval and visualization because of the rich set of tasks that users want to accomplish. This paper demonstrates a tool, NetLens, to explore a Content-Actor paired...
详细信息
ISBN:
(纸本)1595933549
Networks have remained a challenge for information retrieval and visualization because of the rich set of tasks that users want to accomplish. This paper demonstrates a tool, NetLens, to explore a Content-Actor paired network data model. The NetLens interface was designed to allow users to pose a series of elementary queries and iteratively refine visual overviews and sorted lists. This enables the support of complex queries that are traditionally hard to specify in node-link visualizations. NetLens is general and scalable in that it applies to any dataset that can be represented with our abstract Content-Actor data model.
Given a set of radio broadcast programs, the radio broadcast scheduling problem is to allocate a set of devices to transmit the programs to achieve the optimal sound quality. In this article, we propose a complete alg...
详细信息
Given a set of radio broadcast programs, the radio broadcast scheduling problem is to allocate a set of devices to transmit the programs to achieve the optimal sound quality. In this article, we propose a complete algorithm to solve the problem, which is based on a branch-and-bound(BnB) algorithm. We formulate the problem with a new model, called constrained maximum weighted bipartite matching(CMBM),i.e., the maximum matching problem on a weighted bipartite graph with constraints. For the reduced matching problem, we propose a novel BnB algorithm by introducing three new strategies, including the highest quality first, the least conflict first and the more edge first. We also establish an upper bound estimating function for pruning the search space of the algorithm. The experimental results show that our new algorithm can quickly find the optimal solution for the radio broadcast scheduling problem at small scales, and has higher scalability for the problems at large scales than the existing complete algorithm.
iSonic is an interactive sonification tool for vision impaired users to explore geo-referenced statistical data, such as population or crime rates by geographical regions. Users use a keyboard or a smooth surface touc...
详细信息
ISBN:
(纸本)1595931597
iSonic is an interactive sonification tool for vision impaired users to explore geo-referenced statistical data, such as population or crime rates by geographical regions. Users use a keyboard or a smooth surface touchpad to interact with coordinated map and table views of the data. The integrated use of musical sounds and speech allows users to grasp the overall data trends and to explore the data to get more details. Scenarios of use are described.
Gesture based interfaces promise to increase the efficiency of user input, particularly in mobile computing where standard input devices such as the mouse and keyboard are impractical. This paper describes an investig...
详细信息
暂无评论