Using Eye-tracking and Deep Learning Approach to Promote Naturalness in Urban Landscape

2020 ◽  
Author(s):  
Yeshan Qiu ◽  
Yugang Chen ◽  
Shengquan Che
Keyword(s):  
Deep Learning ◽  
Eye Tracking ◽  
Bayesian Network ◽  
Evaluation System ◽  
Public Perception ◽  
Urban Landscape ◽  
Evaluation Model ◽  
Landscape Composition ◽  
Evaluation Framework ◽  
Aesthetic Evaluation

<p>Promoting greenness and naturalness has been the integral goal in nature-based solutions for urban environments. Design and building appreciated landscape for subjective public perception is a key factor in the success of promoting urban greenness and naturalness. The current measures of naturalness are siloed from public appreciation and acceptance of urban landscape designs. Our goal is to use state-of-art methods combining traditional design perception evaluation to embed naturalness with public landscape aesthetic perceptions evaluation system. A deep learning and eye-tracking based approach to understand public aesthetic perceptions of landscape street-view images is developed and applied to a case study of Shanghai. We use machine deep learning techniques to identify and assess landscape composition with landscape images and in-situ captured data to study the influence of naturalness of public perceptions of landscape based on a Bayesian network aesthetic evaluation model. The methodology extend the present landscape aesthetic evaluation framework and has the potential to be implemented to much wider applications. Our results indicate a co-conception of naturalness and public appreciation as a proof-of-concept of nature-based solutions.</p><p>Key words：Eye-tracking；Deep Learning；Naturalness；Public aesthetic perceptions；Bayesian network aesthetic evaluation</p>

scholarly journals Automatic Visual Attention Detection for Mobile Eye Tracking Using Pre-Trained Computer Vision Models and Human Gaze

Sensors ◽  
2021 ◽  
Vol 21 (12) ◽  
pp. 4143
Author(s):  
Michael Barz ◽  
Daniel Sonntag
Keyword(s):  
Deep Learning ◽  
Visual Attention ◽  
Eye Tracking ◽  
Performance Metrics ◽  
Evaluation Framework ◽  
Detection Methods ◽  
Manual Annotation ◽  
Learning Models ◽  
Areas Of Interest ◽  
Mobile Eye Tracking

Processing visual stimuli in a scene is essential for the human brain to make situation-aware decisions. These stimuli, which are prevalent subjects of diagnostic eye tracking studies, are commonly encoded as rectangular areas of interest (AOIs) per frame. Because it is a tedious manual annotation task, the automatic detection and annotation of visual attention to AOIs can accelerate and objectify eye tracking research, in particular for mobile eye tracking with egocentric video feeds. In this work, we implement two methods to automatically detect visual attention to AOIs using pre-trained deep learning models for image classification and object detection. Furthermore, we develop an evaluation framework based on the VISUS dataset and well-known performance metrics from the field of activity recognition. We systematically evaluate our methods within this framework, discuss potentials and limitations, and propose ways to improve the performance of future automatic visual attention detection methods.

A Modular-Based Supplier Evaluation Framework: A Comprehensive Data Analysis of ANP Structure

2014 ◽  
Vol 13 (05) ◽  
pp. 883-916 ◽  
Author(s):  
Daji Ergu ◽  
Gang Kou ◽  
Jennifer Shang
Keyword(s):  
Sensitivity Analysis ◽  
Evaluation System ◽  
Missing Values ◽  
Evaluation Model ◽  
Questionnaire Design ◽  
Evaluation Framework ◽  
Design Matrix ◽  
Consistency Test ◽  
Rank Reversal ◽  
Network Process

The analytical network process (ANP) has been widely used to evaluate the suppliers, an important subject in supply chain management. We propose an integrated ANP-based evaluation model, which systematically examines six key decision-making modules. They are: questionnaire design, matrix classification, consistency test, inconsistency identification, uncertain or missing values estimation, and sensitivity analysis of rank reversal. Module 1 involves questionnaire design formats, and Module 2 classifies the collected data. Module 3 introduces the maximum eigenvalue threshold as a critical value for consistency test. The induced bias matrix model (IBMM) is employed in Modules 4 and 5 to identify the inconsistent elements and to estimate the uncertain or missing values. In Module 6, we extend the concept of the IBMM to check for rank reversal when a new alternative or criterion is added that perturbs the evaluation system. The case company in the healthcare device industry under study validates that the proposed method is effective. The integrated model not only improves the questionnaire design and simplifies the ANP consistency tests, but also effectively identifies the inconsistent elements and estimates the missing values. The innovative approach to sensitivity analysis is especially insightful and contributes to the understanding of rank reversal issue.

scholarly journals An Adversarial Deep-Learning-Based Model for Cervical Cancer CTV Segmentation With Multicenter Blinded Randomized Controlled Validation

2021 ◽  
Vol 11 ◽  
Author(s):  
Zhikai Liu ◽  
Wanqi Chen ◽  
Hui Guan ◽  
Hongnan Zhen ◽  
Jing Shen ◽  
...  
Keyword(s):  
Cervical Cancer ◽  
Deep Learning ◽  
Evaluation System ◽  
Performance Metrics ◽  
Locally Advanced ◽  
Evaluation Framework ◽  
Test Accuracy ◽  
Randomized Controlled ◽  
Ct Data ◽  
Controlled Evaluation

PurposeTo propose a novel deep-learning-based auto-segmentation model for CTV delineation in cervical cancer and to evaluate whether it can perform comparably well to manual delineation by a three-stage multicenter evaluation framework.MethodsAn adversarial deep-learning-based auto-segmentation model was trained and configured for cervical cancer CTV contouring using CT data from 237 patients. Then CT scans of additional 20 consecutive patients with locally advanced cervical cancer were collected to perform a three-stage multicenter randomized controlled evaluation involving nine oncologists from six medical centers. This evaluation system is a combination of objective performance metrics, radiation oncologist assessment, and finally the head-to-head Turing imitation test. Accuracy and effectiveness were evaluated step by step. The intra-observer consistency of each oncologist was also tested.ResultsIn stage-1 evaluation, the mean DSC and the 95HD value of the proposed model were 0.88 and 3.46 mm, respectively. In stage-2, the oncologist grading evaluation showed the majority of AI contours were comparable to the GT contours. The average CTV scores for AI and GT were 2.68 vs. 2.71 in week 0 (P = .206), and 2.62 vs. 2.63 in week 2 (P = .552), with no significant statistical differences. In stage-3, the Turing imitation test showed that the percentage of AI contours, which were judged to be better than GT contours by ≥5 oncologists, was 60.0% in week 0 and 42.5% in week 2. Most oncologists demonstrated good consistency between the 2 weeks (P > 0.05).ConclusionsThe tested AI model was demonstrated to be accurate and comparable to the manual CTV segmentation in cervical cancer patients when assessed by our three-stage evaluation framework.

scholarly journals The Effectiveness of Pictorial Aesthetics Based on Multiview Parallel Neural Networks in Art-Oriented Teaching

2021 ◽  
Vol 2021 ◽  
pp. 1-14
Author(s):  
Liang Gang ◽  
Gao Weishang
Keyword(s):  
Evaluation Model ◽  
Evaluation Framework ◽  
Aesthetic Evaluation ◽  
Learning Techniques ◽  
Art Teaching ◽  
Art Classroom ◽  
Multiple Input ◽  
The Aesthetic ◽  
Global And Local ◽  
Classroom Simulation

How to effectively improve the effectiveness of art teaching has always been one of the hot topics concerned by all sectors of society. Especially, in art teaching, situational interaction helps improve the atmosphere of art class. However, there are few attempts to quantitatively evaluate the aesthetics of ink painting. Ink painting expresses images through ink tone and stroke changes, which is significantly different from photos and paintings in visual characteristics, semantic characteristics, and aesthetic standards. For this reason, this study proposes an adaptive computational aesthetic evaluation framework for ink painting based on situational interaction using deep learning techniques. The framework extracts global and local images as multiple input according to the aesthetic criteria of ink painting and designs a model named MVPD-CNN to extract deep aesthetic features; finally, an adaptive deep aesthetic evaluation model is constructed. The experimental results demonstrate that our model has higher aesthetic evaluation performance compared with baseline, and the extracted deep aesthetic features are significantly better than the traditional manual design features, and its adaptive evaluation results reach a Pearson height of 0.823 compared with the manual aesthetic. In addition, art classroom simulation and interference experiments show that our model is highly resistant to interference and more sensitive to the three painting elements of composition, ink color, and texture in specific compositions.

scholarly journals A Study on Improving Disaster Management Evaluation Based on the PUFA Meta Evaluation Model

2020 ◽  
Vol 20 (1) ◽  
pp. 101-110
Author(s):  
Young Hee Kim ◽  
Soo Jun Kim ◽  
Ji Ho Lee
Keyword(s):  
Disaster Management ◽  
Local Governments ◽  
Evaluation System ◽  
Evaluation Model ◽  
Public Institutions ◽  
Evaluation Framework ◽  
Research Model ◽  
Point Scale ◽  
Management Evaluation ◽  
Disaster Assessment

This study aims to examine the problems of the DME (Disaster Management Evaluation) System, which has been implemented since 2005, using meta evaluation; it also aims topropose improvements. For this purpose, the research model was built on the four standards (propriety, utility, feasibility, accuracy) presented in the PUFA meta evaluation model. Checklist items corresponding to the PUFA model sub-components were then selected, and questionnaires were created using a Likert-style5-point scale for disaster assessment evaluators, central departments, local governments, and public institutions. Some of the results of thePUFA sub-components were statistically significant and proved to be a useful meta-evaluation framework for assessing the DME level. For evaluation performance, DME can only be improved if policy efforts to ensure the expertise of the evaluation department, credibility of the evaluator, and justification for the evaluation results are continued. From an evaluation indicators perspective, there is need for improvement ofthe capability-performance multi-dimensional evaluation system so that it can take into account the unique enforcement functions and regional characteristics of the evaluation target organization.

Fuzzy Comprehensive Evaluation of Human Performance Management

2013 ◽  
Vol 756-759 ◽  
pp. 715-719
Author(s):  
Huan Cheng Zhang ◽  
Ya Feng Yang ◽  
Feng Li ◽  
Li Nan Shi
Keyword(s):  
Performance Management ◽  
Evaluation System ◽  
Human Performance ◽  
Comprehensive Evaluation ◽  
Evaluation Model ◽  
Fuzzy Comprehensive Evaluation ◽  
Assessment Process ◽  
College Performance ◽  
Performance Evaluation System ◽  
Comprehensive Evaluation Model

In the College, performance evaluation system is directly related to the harmonious development of the school. Taking into account the factors in the evaluation system is fuzzy, so this paper uses fuzzy comprehensive evaluation model. But the model is too subjective, so this paper combines neural network and data envelopment analysis method, which ensures that fuzzy comprehensive evaluation model is reasonable and scientific, and good school development and teacher self-interest. The performance assessment process, not only enables the combination of qualitative and quantitative analysis, but also fair and reasonably reflect the achievements of teachers, while this method is easy to use, wide application, and can be well applied in practice.

scholarly journals Evaluation Model of Operation State Based on Deep Learning for Smart Meter

Energies ◽  
2021 ◽  
Vol 14 (15) ◽  
pp. 4674
Author(s):  
Qingsheng Zhao ◽  
Juwen Mu ◽  
Xiaoqing Han ◽  
Dingkang Liang ◽  
Xuping Wang
Keyword(s):  
Deep Learning ◽  
Power Consumption ◽  
Detection Efficiency ◽  
Evaluation Model ◽  
Smart Meter ◽  
Smart Meters ◽  
Abnormal Detection ◽  
Prediction Residual ◽  
Threshold Setting ◽  
Training Efficiency

The operation state detection of numerous smart meters is a significant problem caused by manual on-site testing. This paper addresses the problem of improving the malfunction detection efficiency of smart meters using deep learning and proposes a novel evaluation model of operation state for smart meter. This evaluation model adopts recurrent neural networks (RNN) to predict power consumption. According to the prediction residual between predicted power consumption and the observed power consumption, the malfunctioning smart meter is detected. The training efficiency for the prediction model is improved by using transfer learning (TL). This evaluation uses an accumulator algorithm and threshold setting with flexibility for abnormal detection. In the simulation experiment, the detection principle is demonstrated to improve efficient replacement and extend the average using time of smart meters. The effectiveness of the evaluation model was verified on the actual station dataset. It has accurately detected the operation state of smart meters.

Study on Civil Architecture Energy Conservation Efficiency Fuzzy Synthetic Evaluation

2011 ◽  
Vol 94-96 ◽  
pp. 2292-2296
Author(s):  
Yong Ping Wang ◽  
Pei Yang ◽  
Chun Quan Dai
Keyword(s):  
Energy Conservation ◽  
Evaluation System ◽  
Evaluation Model ◽  
Evaluation Methodology ◽  
Efficiency Evaluation ◽  
Fuzzy Synthetic Evaluation ◽  
Conservation Efficiency ◽  
Synthetic Evaluation

Civil architecture energy conservation efficiency evaluation is a kind of multi-factors, multi-hierarchies and multi-criteria synthetic evaluation. Perfect civil architecture energy conservation efficiency evaluation indicators system and reasonably effective synthetic evaluation methodology are keys to do energy conservation efficiency synthetic evaluation. This paper is based on framing civil architecture energy conservation efficiency evaluation system, and uses fuzzy synthetic evaluation methodology to frame civil architecture energy conservation efficiency fuzzy synthetic evaluation model, in order to make the result of evaluation more objective and reasonable.

Development of an Evaluation Model of Freeway Traffic Safety Facilities System

2013 ◽  
Vol 734-737 ◽  
pp. 1578-1581
Author(s):  
Yan Yong Guo ◽  
Yao Wu ◽  
Liang Song ◽  
Hui Duan
Keyword(s):  
Traffic Safety ◽  
Evaluation System ◽  
Evaluation Model ◽  
Variation Coefficient ◽  
Freeway Traffic ◽  
Evaluation Result ◽  
Single Index ◽  
Recognition Theory ◽  
Attribute Measure ◽  
Attribute Recognition

This study developed an evaluation model of freeway traffic safety facilities system. Firstly, an evaluation system of freeway traffic safety facility was proposed. Secondly, an evaluation model was proposed based on attribute recognition theory. And the evaluation result was identified according to the attribute measure value of single index and the comprehensive attribute measure value of multiple indexes as well as the confidence criterion. Thirdly, the weight of each indicator was decided by variation coefficient. Finally, A case of TAI-GAN freeway (K1+242~K3+259 segment) was conducted to verify the feasibility and effectiveness of the model.

Sign in / Sign up

Export Citation Format

Share Document