FEATURE FUSION AND MODEL SELECTION BASED ON INFORMATION CRITERION

2006 ◽  
Vol 03 (02) ◽  
pp. 85-99
Author(s):  
TAKAMI SATONAKA ◽  
KEIICHI UCHIMURA
Keyword(s):  
Variance Estimation ◽  
Feature Fusion ◽  
Likelihood Function ◽  
Asymptotic Variance ◽  
Metric Learning ◽  
Information Criterion ◽  
Trade Offs ◽  
Training Samples ◽  
Fusion Methods ◽  
The Face

We describe the k-NN adaptive metric learning procedure combining the asymptotic variance estimation and fine adjustment of the metric parameters for the face recognition. The metric learning model based on the Mahalanobis distance suffered from the degraded performance due to the limitation of available training samples. The feature fusion methods are proposed to assume local distributions of feature patterns for the parameter estimation and learning. Firstly, the MDL criterion is formulated to decide on the trade-offs between accuracy and complexity of an asymptotic statistical model. The variance within the classes is minimized by using the asymptotic variance estimation. Secondly, optimal metric parameters are derived from the minimization of the negative log-likelihood function for the presentation of the synthesized feature patterns. The variance between the classes is increased by using the simulated annealing method. We present the simulation results using the ORL and UMIST databases.

scholarly journals Predicting Implicit User Preferences with Multimodal Feature Fusion for Similar User Recommendation in Social Media

2021 ◽  
Vol 11 (3) ◽  
pp. 1064
Author(s):  
Jenq-Haur Wang ◽  
Yen-Tsang Wu ◽  
Long Wang
Keyword(s):  
Social Media ◽  
Feature Fusion ◽  
Relevant Information ◽  
Image Features ◽  
User Preferences ◽  
User Preference ◽  
Late Fusion ◽  
Multimodal Features ◽  
Fusion Methods ◽  
Text Features

In social networks, users can easily share information and express their opinions. Given the huge amount of data posted by many users, it is difficult to search for relevant information. In addition to individual posts, it would be useful if we can recommend groups of people with similar interests. Past studies on user preference learning focused on single-modal features such as review contents or demographic information of users. However, such information is usually not easy to obtain in most social media without explicit user feedback. In this paper, we propose a multimodal feature fusion approach to implicit user preference prediction which combines text and image features from user posts for recommending similar users in social media. First, we use the convolutional neural network (CNN) and TextCNN models to extract image and text features, respectively. Then, these features are combined using early and late fusion methods as a representation of user preferences. Lastly, a list of users with the most similar preferences are recommended. The experimental results on real-world Instagram data show that the best performance can be achieved when we apply late fusion of individual classification results for images and texts, with the best average top-k accuracy of 0.491. This validates the effectiveness of utilizing deep learning methods for fusing multimodal features to represent social user preferences. Further investigation is needed to verify the performance in different types of social media.

scholarly journals Research on a Face Real-time Tracking Algorithm Based on Particle Filter Multi-Feature Fusion

Sensors ◽  
2019 ◽  
Vol 19 (5) ◽  
pp. 1245 ◽  
Author(s):  
Tao Wang ◽  
Wen Wang ◽  
Hui Liu ◽  
Tianping Li
Keyword(s):  
Big Data ◽  
Particle Filter ◽  
Skin Color ◽  
Feature Fusion ◽  
Average Distance ◽  
Histogram Method ◽  
Face Features ◽  
The Face ◽  
Edge Features ◽  
Tracking Object

With the revolutionary development of cloud computing and internet of things, the integration and utilization of “big data” resources is a hot topic of the artificial intelligence research. Face recognition technology information has the advantages of being non-replicable, non-stealing, simple and intuitive. Video face tracking in the context of big data has become an important research hotspot in the field of information security. In this paper, a multi-feature fusion adaptive adjustment target tracking window and an adaptive update template particle filter tracking framework algorithm are proposed. Firstly, the skin color and edge features of the face are extracted in the video sequence. The weighted color histogram are extracted which describes the face features. Then we use the integral histogram method to simplify the histogram calculation of the particles. Finally, according to the change of the average distance, the tracking window is adjusted to accurately track the tracking object. At the same time, the algorithm can adaptively update the tracking template which improves the accuracy and accuracy of the tracking. The experimental results show that the proposed method improves the tracking effect and has strong robustness in complex backgrounds such as skin color, illumination changes and face occlusion.

Advances in Fisheries Bioengineering

2008 ◽  
Author(s):  
Keyword(s):  
River Flow ◽  
Oncorhynchus Tshawytscha ◽  
Prolonged Exposure ◽  
Likelihood Function ◽  
Predictor Variable ◽  
Passage Time ◽  
Fish Length ◽  
River Temperature ◽  
Relative Contribution ◽  
The Face

<em>Abstract.</em>—Although many hydroelectric dams have fishways for upstream passage of migratory fish, passage delays often occur at these sites. Migrational delay may affect fish detrimentally in several ways, including depletion of energy reserves, suboptimal arrival timing at spawning sites, and prolonged exposure to hazardous conditions at the face of dams. We applied time-to-event analyses to passage times of radio-tagged adult Chinook salmon <em>Oncorhynchus tshawytscha </em>at four dams on the lower Columbia River, where many fish require several days to pass each dam. The analysis allowed us to determine instantaneous passage rates in response to fluctuating river conditions. By relating variability in passage rate to the predictor variables river temperature, river flow, and fish size, we determined the relative contribution of various factors to the passage time of migrating fish. We fit the model by maximizing the likelihood function that incorporated information from individuals rather than aggregated groups of fish. We used Akaike’s Information Criterion to distinguish among several competing models, each of which used a different predictor variable. We found that daytime passage rates were significantly greater than nighttime passage rates. Also, the influence of river flow, river temperature, and fish length on passage rates varied at the four dams. However, when a factor had a significant influence on passage time, the direction of the relationship was consistent across dams: river flow and fish length were positively related to passage time (greater values led to longer passage time), and river temperature was negatively related. This method is easily adaptable to study passage time of any fish population facing a broad range of obstacles to migration, whether natural or man-made.

Analysis of Sustainable Development Path for Green Finance

2019 ◽  
pp. 62-83
Author(s):  
Hsu Chao Feng ◽  
Lee Bi Ru
Keyword(s):  
Sustainable Development ◽  
Environmental Protection ◽  
National Development ◽  
Development Project ◽  
Market Society ◽  
Rapid Economic Growth ◽  
Trade Offs ◽  
Green Finance ◽  
The Face ◽  
The Government

The development of green finance is a global trend in the current era. At present, developing the green finance has been included as an important national development project by the Chinese government. With the rapid economic growth, the priorities or trade-offs between the economic development and the natural environment have also aroused different contradictions and problems. With the improvement of people's quality of life, they start to pay more attention to the pollution of the surrounding environment. Therefore, the government should properly intervene and propose effective measures, and green finance is an excellent tool to reconcile social economy and environmental protection and transform the physical investment, thus guiding the social resources towards the environmental protection industry and reaching an optimal interests allocation among the market, society, and government. Consequently, in the face of such a situation, it is necessary to propose a series of models and paths that suit the needs of the Chinese society and promote sustainable development.

scholarly journals DTFA-Net: Dynamic and Texture Features Fusion Attention Network for Face Antispoofing

Complexity ◽  
2020 ◽  
Vol 2020 ◽  
pp. 1-11
Author(s):  
Xin Cheng ◽  
Hongfei Wang ◽  
Jingmei Zhou ◽  
Hui Chang ◽  
Xiangmo Zhao ◽  
...  
Keyword(s):  
Face Recognition ◽  
Face Detection ◽  
Feature Fusion ◽  
Texture Features ◽  
Replay Attack ◽  
Attention Network ◽  
Detection Failure ◽  
The Face ◽  
Recognition Systems ◽  
Common Face

For face recognition systems, liveness detection can effectively avoid illegal fraud and improve the safety of face recognition systems. Common face attacks include photo printing and video replay attacks. This paper studied the differences between photos, videos, and real faces in static texture and motion information and proposed a living detection structure based on feature fusion and attention mechanism, Dynamic and Texture Fusion Attention Network (DTFA-Net). We proposed a dynamic information fusion structure of an interchannel attention block to fuse the magnitude and direction of optical flow to extract facial motion features. In addition, for the face detection failure of HOG algorithm under complex illumination, we proposed an improved Gamma image preprocessing algorithm, which effectively improved the face detection ability. We conducted experiments on the CASIA-MFSD and Replay Attack Databases. According to experiments, the DTFA-Net proposed in this paper achieved 6.9% EER on CASIA and 2.2% HTER on Replay Attack that was comparable to other methods.

scholarly journals Sustainability in the Canadian Egg Industry—Learning from the Past, Navigating the Present, Planning for the Future

2018 ◽  
Vol 10 (10) ◽  
pp. 3524 ◽  
Author(s):  
Nathan Pelletier ◽  
Maurice Doyon ◽  
Bruce Muirhead ◽  
Tina Widowski ◽  
Jodey Nurse-Gupta ◽  
...  
Keyword(s):  
Animal Welfare ◽  
Value Chain ◽  
Housing Systems ◽  
The Past ◽  
Trade Offs ◽  
The Face ◽  
Effective Decision Making ◽  
And Shifts ◽  
Societal Expectations ◽  
Management Changes

Like other livestock sectors, the Canadian egg industry has evolved substantially over time and will likely experience similarly significant change looking forward, with many of these changes determining the sustainability implications of and for the industry. Influencing factors include: technological and management changes at farm level and along the value chain resulting in greater production efficiencies and improved life cycle resource efficiency and environmental performance; a changing policy/regulatory environment; and shifts in societal expectations and associated market dynamics, including increased attention to animal welfare outcomes—especially in regard to changes in housing systems for laying hens. In the face of this change, effective decision-making is needed to ensure the sustainability of the Canadian egg industry. Attention both to lessons from the past and to the emerging challenges that will shape its future is required and multi- and interdisciplinary perspectives are needed to understand synergies and potential trade-offs between alternative courses of action across multiple aspects of sustainability. Here, we consider the past, present and potential futures for this industry through the lenses of environmental, institutional (i.e., regulatory), and socio-economic sustainability, with an emphasis on animal welfare as an important emergent social consideration. Our analysis identifies preferred pathways, potential pitfalls, and outstanding cross-disciplinary research questions.

scholarly journals Efficiency and thoroughness trade-offs in high-volume organisational routines: an ethnographic study of prescribing safety in primary care

2017 ◽  
Vol 27 (3) ◽  
pp. 199-206 ◽  
Author(s):  
Suzanne Grant ◽  
Bruce Guthrie
Keyword(s):  
Primary Care ◽  
Participant Observation ◽  
High Volume ◽  
Ethnographic Study ◽  
Trade Offs ◽  
High Workload ◽  
Organisational Routines ◽  
The Face ◽  
General Practices ◽  
Context Specific

BackgroundPrescribing is a high-volume primary care routine where both speed and attention to detail are required. One approach to examining how organisations approach quality and safety in the face of high workloads is Hollnagel’s Efficiency and Thoroughness Trade-Off (ETTO). Hollnagel argues that safety is aligned with thoroughness and that a choice is required between efficiency and thoroughness as it is not usually possible to maximise both. This study aimed to ethnographically examine the efficiency and thoroughness trade-offs made by different UK general practices in the achievement of prescribing safety.MethodsNon-participant observation was conducted of prescribing routines across eight purposively sampled UK general practices. Sixty-two semistructured interviews were also conducted with key practice staff alongside the analysis of relevant practice documents.ResultsThe eight practices in this study adopted different context-specific approaches to safely handling prescription requests by variably prioritising speed of processing by receptionists (efficiency) or general practitioner (GP) clinical judgement (thoroughness). While it was not possible to maximise both at the same time, practices situated themselves at various points on an efficiency-thoroughness spectrum where one approach was prioritised at particular stages of the routine. Both approaches carried strengths and risks, with thoroughness-focused approaches considered safer but more challenging to implement in practice due to GP workload issues. Most practices adopting efficiency-focused approaches did so out of necessity as a result of their high workload due to their patient population (eg, older, socioeconomically deprived).ConclusionsHollnagel’s ETTO presents a useful way for healthcare organisations to optimise their own high-volume processes through reflection on where they currently prioritise efficiency and thoroughness, the stages that are particularly risky and improved ways of balancing competing priorities.

Catastrophic Climate Change, Precaution, and the Risk/Risk Dilemma

2017 ◽  
Author(s):  
Floor M Fleurke
Keyword(s):  
Climate Change ◽  
Precautionary Principle ◽  
New Technology ◽  
Targeted Intervention ◽  
Regulatory Response ◽  
The Precautionary Principle ◽  
Trade Offs ◽  
The Face

Whilst the seriousness of a given problem may call for immediate and targeted intervention, the ensuing uncertain impacts on other elements of inter-connected systems may be equally deleterious. Climate change is a prime example of such a risk/risk dilemma. The risk of inaction must be weighed against the risk of resorting to increasingly tempting responses to mitigate or adapt to the effects of climate change. The precautionary principle might offer some guidance in this risk/risk arena. Precaution is a tool to deal with uncertain risks without dictating outcomes. Although it is commonly associated with a negative regulatory tilt, it can also serve to warrant and mandate the use of, for example, a new technology or substance in order to reduce risks. This chapter explores the dilemma of risk/risk trade-offs in the face of potentially catastrophic climate change, and examines the contours of a precautionary regulatory response to such impasses.

Sign in / Sign up

Export Citation Format

Share Document