This comprehensive review paper explores the state of the art in visual localization and navigation, drawing on the principles and methodologies of three significant deep learning networks: visual Localization Network...
This comprehensive review paper explores the state of the art in visual localization and navigation, drawing on the principles and methodologies of three significant deep learning networks: visual Localization Network (vLocNet), Deep Fusion Network (DFNet), and Hybrid Frontend Network (HFNet). Each of these networks demonstrates the application of deep learning to spatial awareness and navigation tasks in unique and significant ways. Rather than an exhaustive dissection of these networks, the paper provides an encompassing overview, illuminating their underlying principles, architectural design, and their relative performance within the field. Additionally, the paper delves into the practical implications of these networks, examining their applications in diverse real-world scenarios. It underlines this examination with a comprehensive analysis of existing literature and experimental results, intended to impart a profound understanding of these networks' strengths, limitations, and potential application areas. Ultimately, this review aims to present a valuable compass to researchers navigating the evolving landscape of advancements in visual localization and navigation, thereby fostering enriched understanding and facilitating future exploration and development in this compelling field.
This paper presents a solution for the autonomous exploration and inspection of Ballast Water Tanks (BWTs) of marine vessels using aerial robots. Ballast tank compartments are critical for a vessel's safety and co...
This paper presents a solution for the autonomous exploration and inspection of Ballast Water Tanks (BWTs) of marine vessels using aerial robots. Ballast tank compartments are critical for a vessel's safety and correspond to confined environments often connected through particularly narrow manholes. The method enables their volumetric exploration combined with visual inspection subject to constraints regarding the viewing distance from a surface. We present evaluation studies in simulation, in a mission consisting of 18 BWT compartments, and in 3 field experiments inside real vessels. The data from one of the experiments is also post-processed to generate semantically-segmented meshes of inspection-important geometries. Geometric models can be associated with onboard camera images for detailed and intuitive analysis.
The current research on acoustic properties in singing voice analysis has mainly focused on individual song segments, and analysed them using simple data tables and basic charts. However, there has been limited explor...
The current research on acoustic properties in singing voice analysis has mainly focused on individual song segments, and analysed them using simple data tables and basic charts. However, there has been limited exploration of comparing data from multiple sources, and visualanalysis in crossover singing has been either too simplistic or too complex to provide a comprehensive view. This study aims to address this gap by incorporating song segments from different musical styles and utilising an innovative graph drawing method to generate interactive graphs for comprehensive musical dataanalysis. The findings provide additional support for existing statements in the field of musical dataanalysis, and demonstrate the effectiveness of the proposed graph method for analysing multiple song segments. At this stage, the study's findings confirm that the formant frequency $F_exploration$ of singing across styles is less modified, but $F_analysis-F_analysis$ varies in styles singing in English. Additionally, the formant frequency in Mandarin and Cantonese singing may be associated with pitch. The study also identifies that visualised graphs can produce similar results as current vocal research and are convenient for reading multiple data simultaneously. The methodology has the potential to be extended to other areas of musical visualisation to uncover insights from complex music datasets.
A considerable amount of useful information on the web is (semi-)structured, such as tables and lists. An extensive corpus of prior work addresses the problem of making these human-readable representations interpretab...
详细信息
ISBN:
(纸本)9781728191843
A considerable amount of useful information on the web is (semi-)structured, such as tables and lists. An extensive corpus of prior work addresses the problem of making these human-readable representations interpretable by algorithms. Most of these works focus only on the most recent snapshot of these web objects. However, their evolution over time represents valuable information that has barely been tapped, enabling various applications, including visual change exploration and trust assessment. To realize the full potential of this information, it is critical to match such objects across page revisions. In this work, we present novel techniques that match tables, infoboxes and lists within a page across page revisions. We are, thus, able to extract the evolution of structured information in various forms from a long series of web page revisions. We evaluate our approach on a representative sample of pages and measure the number of correct matches. Our approach achieves a significant improvement in object matching over baselines and over related work.
This study aimed to provide greater insight into the question of whether a near-field Paleozoic interval could retain sufficient reservoir quality to be attractive for future exploration. To meet this challenge, an in...
详细信息
The categorization of brain tumors is crucial for accurate medical analysis as well as healing. Convolutional Neural Network plays an essential role in diagnosing disease in the domain of deep learning algorithms whic...
详细信息
Most of western countries have successfully deployed smart meters and are now in a phase of exploration to regulate and leverage the use of smart meter data. In France, the Linky smart meter provides data locally to t...
详细信息
With the development of the times, the use of 3D image virtual reconstruction (IvR) systems is increasing. The 3D IvR system can solve the problems of poor user experience quality and low human-machine interaction eff...
With the development of the times, the use of 3D image virtual reconstruction (IvR) systems is increasing. The 3D IvR system can solve the problems of poor user experience quality and low human-machine interaction efficiency, so the research significance is very significant. However, nowadays, 3D IvR systems are facing issues of reconstruction accuracy. If this problem can be solved, it can greatly improve the efficiency of the 3D IvR system. Therefore, this article focused on the analysis of a 3D IvR system based on visual communication technology, aiming to improve the reconstruction accuracy of the 3D IvR system through visual communication technology. This article experimentally tested the reconstruction accuracy of a 3D IvR system using visual communication technology, with a maximum accuracy of 74% and a minimum accuracy of 62%. However, traditional 3D IvR systems had a maximum accuracy of 56% and a minimum accuracy of 50%. From this experimental data, it can be inferred that the use of visual communication technology can improve the reconstruction accuracy of 3D IvR systems. This indicated that visual communication technology has achieved good results in 3D IvR systems.
Modern earth sciences attach great importance to three-dimensional modeling for the analysis of geological objects of varying complexity and geological and geophysical content. In the field of computer graphics, 3D mo...
详细信息
Planting by mounding is a commonly used forestry technique that improves soil quality and ensures optimal tree growth conditions. During planting operations, one of the main planning steps is to estimate the number of...
详细信息
ISBN:
(纸本)9798350302493
Planting by mounding is a commonly used forestry technique that improves soil quality and ensures optimal tree growth conditions. During planting operations, one of the main planning steps is to estimate the number of mechanically created mounds in each planting block. Traditional counting methods involve manual field surveys or human photo-interpretation of UAv images, which are generally subject to errors and time-consuming. In this work, we propose a new approach to count mounds on UAv orthomosaics. Our framework is designed to estimate the required number of seedlings for a given planting block, based on a visual detection approach and a global estimation module. Firstly, a deep local detection model is applied on local patches to recognize and count visible mounds. Then, an estimation model, based on global features is used to predict the final number of plant seedling required for a given plantation block. To evaluate the proposed framework in real-world conditions, we constructed a large UAvdataset, including 18 UAv orthomosaics, comprising 111,000 mounds. We have conducted extensive experiments in our dataset, including a comparison with the state-of-the-art counting methods, as well as an analysis of Human-Level Performance (HLP) in identifying and annotating mounds. The experimental results show that our model reaches the best performance in terms of MAE and MSE, by comparison to state-of-the-art automatic counting mehtods.
暂无评论