Deep learning-based models have been shown to outperform human beings in many computer vision tasks with massive available labeled training data in learning. However, humans have an amazing ability to easily recognize...
详细信息
Deep learning-based models have been shown to outperform human beings in many computer vision tasks with massive available labeled training data in learning. However, humans have an amazing ability to easily recognize images of novel categories by browsing only a few examples of these categories. In this case, few-shot learning comes into being to make machines learn from extremely limited labeled examples. One possible reason why human beings can well learn novel concepts quickly and efficiently is that they have sufficient visual and semantic prior knowledge. Toward this end, this work proposes a novel knowledge-guided semantic transfer network (KSTNet) for few-shot image recognition from a supplementary perspective by introducing auxiliary prior knowledge. The proposed network jointly incorporates vision inferring, knowledge transferring, and classifier learning into one unified framework for optimal compatibility. A category-guided visual learning module is developed in which a visual classifier is learned based on the feature extractor along with the cosine similarity and contrastive loss optimization. To fully explore prior knowledge of category correlations, a knowledge transfer network is then developed to propagate knowledge information among all categories to learn the semantic-visual mapping, thus inferring a knowledge-based classifier for novel categories from base categories. Finally, we design an adaptive fusion scheme to infer the desired classifiers by effectively integrating the above knowledge and visual information. Extensive experiments are conducted on two widely used Mini-ImageNet and Tiered-ImageNet benchmarks to validate the effectiveness of KSTNet. Compared with the state of the art, the results show that the proposed method achieves favorable performance with minimal bells and whistles, especially in the case of one-shot learning.
Exploratory dataanalysis provides visualexploration of data and also allows for the identification of trends and relationships between variables. The aim of this research is to study about different factors involved...
详细信息
Recognition of human's emotional states is one of crucial issues in many Brain-Computer-Interface (BCI) applications. In the current study, the problem of emotion recognition based on EEG data will be addressed. T...
详细信息
The on-going evolution of technology of capturing and generation video, sound, RF, etc. signals results in very HDR data to be stored and processed. The analysis of human perception of different signals is showing tha...
详细信息
In this paper we present a virtual Reality (vR) tool that facilitates the visualisation and exploration of context-based, multi-level annotations in archaeology. Indeed, thanks to photogrammetry and laser scanning tec...
详细信息
This study presents an innovative approach to emotion recognition by integrating RGB Kinect video data with physiological signals, including electroencephalography (EEG), electrocardiography (ECG), and galvanic skin r...
详细信息
Public databases such as the NCBI Gene Expression Omnibus (GEO) house millions of experimental gene expression datasets invaluable for transcriptome meta-analysis, enabling researchers to identify genes, pathways, and...
详细信息
The growth of cities calls for regulations and zoning rules on how each piece of urban space will be used. Tracking land use can reveal a wealth of information about urban development. For that matter, cities have bee...
详细信息
ISBN:
(纸本)9798350338737;9798350338720
The growth of cities calls for regulations and zoning rules on how each piece of urban space will be used. Tracking land use can reveal a wealth of information about urban development. For that matter, cities have been releasing data sets describing the historical evolution of the shape and the attributes of land units. The complex nature of land-use data, however, makes the analysis of such data challenging and time-consuming. To address these challenges, we propose URBAN CHRONICLES, a visual analytics system that enables interactive exploration of land-use changes. Using New York City's Primary Land Use Tax Lot Output (PLUTO), we show the system's capabilities to explore the data from several years at different scales. URBAN CHRONICLES supports on-the-fly aggregation and filtering operations that leverage the hierarchical nature of the data set to index the shape and attributes of geographical regions that change over time. Finally, we demonstrate the system's utility through case studies that analyze the impact of Hurricane Sandy on land use attributes and the effects of rezoning plans in Brooklyn.
The leather industry confronts persistent challenges in the maintenance of its product quality, often produced with defects that escape from the traditional conventional inspection methods. This research seeks the tra...
详细信息
In this paper, we address the challenges in achieving sustainable data-driven efficiency by providing a detailed exploration of the end-to-end operational data analytics (ODA) framework that evolved through two genera...
详细信息
暂无评论