Urban Classification Research Articles

ResNet Based on Multi-Feature Attention Mechanism for Sound Classification in Noisy Environments

Environmental noise affects people’s lives and poses challenges for urban sound classification. Traditional algorithms such as Mel frequency cepstral coefficients (MFCCs) struggle due to audio signal complexity. This study applied an attention mechanism to a deep residual network (ResNet) deep learning network to overcome the structural impact of urban noise on audio signals and improve classification accuracy. We propose a three-feature fusion ResNet + attention method (Net50_SE) to maximize information representation in environmental sound signals. This method uses residual structured convolutional neural networks (CNNs) for feature extraction in sound classification tasks. Additionally, an attention module is added to suppress environmental noise impact and focus on different feature map channels. The experimental results demonstrate the effectiveness of our method, achieving 93.2% accuracy compared with 82.87% with CNN and 84.77% with long short-term memory (LSTM). Our model provides higher accuracy and confidence in urban sound classification.

Identification and Classification of Urban Shrinkage in Northeast China

The phenomenon of shrinking cities is a significant challenge faced by many cities today. To more accurately identify the leading factors driving urban shrinkage and develop rational recommendations, precise identification and classification of urban shrinkage has become an indispensable part of the process. This paper focuses on the typical population loss region of China’s three northeastern provinces, using 497 identified physical cities as the basic research unit. Based on multi-source geographical big data and utilizing the geographically weighted regression (GWR) model, spatial modeling of population in the three provinces of northeast China was conducted, resulting in spatialized population data, followed by identification and classification of shrinking cities among the physical cities. Cities with a total population change rate of less than 0 are defined as shrinking cities. In cities where the total population change rate is greater than 0, cities with both a city shrinking area ratio and a decreased population ratio greater than 5% are defined as locally shrinking cities. Based on this, 90 (18.1%) shrinking cities and 118 (23.7%) locally shrinking cities were identified within the three provinces of northeast China. The phenomenon of urban shrinkage is distributed throughout various regions, mainly in smaller cities located near larger cities. According to the standards of the urban shrinkage classification model, the spatial pattern of population loss regions was divided into four types, identifying 13 (6.3%) global type, 111 (53.4%) concentrated type, 64 (30.7%) perforated type, and 20 (9.6%) edge type. Analysis of shrinking cities based on their classification revealed that the main reasons for urban shrinkage are the decline and dissolution of large industrial enterprises, abandonment and neglect of buildings, and unreasonable design planning in cities. Economic development and inward population flow can be promoted in shrinking cities by creating job opportunities, improving living standards, developing transportation, adjusting urban planning or concentrating urban population, as well as vigorously developing urban center areas. These measures can provide support for the revival and development of shrinking cities.

Developing Comprehensive Local Climate Zone Land Use Datasets for Advanced High-Resolution Urban Climate and Environmental Modeling

The Local Climate Zone (LCZ) classification scheme is a vital method of building a category dataset for high-resolution urban land. For the development of urban meteorology, air pollution and related disciplines, the high-resolution classification data of urban buildings are very important. This study aims to create LCZ datasets with detailed architectural characteristics for major cities and urban agglomerations in China, and obtain more accurate results. We constructed 120 m resolution land use datasets for 63 cities (mainly provincial capitals, municipalities directly under the Central Government, important prefecture-level cities and special administrative regions) and 4 urban agglomerations in China based on the local climate zone (LCZ) classification scheme using the World Urban Database and Access Portal Tools method (WUDAPT). Nearly 100,000 samples were used, of which 76,000 training samples were used to provide spectral signatures and 23,000 validation samples were used to ensure accuracy assessments. Compared with similar studies, the LCZ datasets in this paper were generally of good quality, with an overall accuracy of 71–93% (mean 82%), an accuracy for built classifications of 57–83% (mean 72%), and an accuracy for natural classifications of 70–99% (mean 90%). In addition, 35% of 63 Chinese cities have construction areas of more than 5%, and the plateaus northwest of Chengdu and Chongqing are covered with snow all year round. Therefore, based on the original LCZ classification system, the construction area (LZC H) and the snow cover (LCZ I) were newly added as the basic classifications of urban LCZ classification in China. Detailed architectural features of cities and urban agglomerations in China are provided by the LCZ datasets in this study. It can be applied to fine numerical models of the meteorological and atmospheric environment and improve the prediction accuracy.

Inflammatory Bowel Disease in Appalachian Kentucky: An Investigation of Outcomes and Health Care Utilization.

Rural residence has been associated with a lower incidence of inflammatory bowel disease (IBD) but higher health care utilization and worse outcomes. Socioeconomic status is intrinsically tied to both IBD incidence and outcomes. Inflammatory bowel disease outcomes have not been investigated in Appalachia: a rural, economically distressed region rife with risk factors for both increased incidence and unfavorable outcomes. Hospital inpatient discharge and outpatient services databases were utilized to assess outcomes in patients diagnosed with either Crohn's disease (CD) or ulcerative colitis (UC) in Kentucky. Encounters were classified by patient residence in Appalachian or non-Appalachian counties. Data were reported as crude and age-adjusted rates of visits per 100,000 population per year collected in 2016 to 2019. National inpatient discharge data from 2019, stratified by rural and urban classification codes, were utilized to compare Kentucky to national trends. Crude and age-adjusted rates of inpatient, emergency department and outpatient encounters were higher in the Appalachian cohort for all 4 years observed. Appalachian inpatient encounters are more frequently associated with a surgical procedure (Appalachian, 676, 24.7% vs non-Appalachian, 1408, 22.2%; P = .0091). In 2019, the Kentucky Appalachian cohort had significantly higher crude and age-adjusted rates of inpatient discharges for all IBD diagnoses compared with national rural and nonrural populations (crude 55.2; 95% CI, 50.9-59.5; age-adjusted 56.7; 95% CI, 52.1-61.3). There is disproportionately higher IBD health care utilization in Appalachian Kentucky compared with all cohorts, including the national rural population. There is a need for aggressive investigation into root causes of these disparate outcomes and identification of barriers to appropriate IBD care.

Polarimetric Decomposition and Machine Learning‐Based Classification of L‐ and S‐Band Airborne SAR (LS‐ASAR) Data

AbstractThe polarimetric Synthetic Aperture Radar (SAR) data sets have been widely exploited for land use land cover (LULC) classification due to their sensitivity to the structural and dielectric properties of the imaging target. In this study, the potential of fully polarimetric L‐ and S‐band Airborne SAR (LS‐ASAR) data sets were explored for the machine‐learning‐based classification of Urban, Vegetation, Waterbody, and Open Ground. This work was done by utilizing dual‐frequency L‐ and S‐band airborne data of Santa Barbara, California, USA, acquired under the Airborne SAR (LS ASAR) campaign, a precursor airborne mission to the space‐borne NASA‐ISRO (NISAR) mission. The LS‐ASAR polarimetric information was utilized for LULC classification using the SVM classifier. The roll‐invariant Barnes, and eigenvalue/eigenvector‐based Cloude and H/A/Alpha decomposition were implemented to retrieve the scattering parameters. The backscatter response of classes was studied, and separability analysis was done to reduce the misclassification error between six class pairs‐ Vegetation—Urban, Vegetation—Waterbody, Vegetation—Open Ground, Urban—Waterbody, Urban—Open Ground, and Water—Open Ground. The decomposition models failed to achieve the desirable separability index for all six class pairs; consequently, the classification of Barnes, Cloude, H/A/Alpha decomposition showed misclassification between vegetation‐urban class, and waterbody‐open ground class for both L‐ and S‐band data sets. The effort was made toward improving the classification accuracy by integrating the roll‐invariant and eigenvalue‐eigenvector scattering parameters of the multifrequency L‐ and S‐band data set. This method presented the desirable separability index for all class‐pair; eventually highest classification accuracy was achieved i.e. 93.35% ( = 0.91) by significantly reducing the misclassification error between class pairs.

Hyperspectral image-aided LiDAR point cloud labeling via spatio-spectral feature representation learning

Urban scene-level 3D point cloud labeling is a very laborious and expensive task compared to images. Conversely however, image processing techniques, deep learning or otherwise are more established and mature. Thus, in a multi-source data environment, the labeling of a point cloud scene via an automated image process as an initial step, followed by a manual human verification process is an effective way to save man hours and cost. With the above as the goal, this study presents a simple but robust spatio-spectral feature representation approach. In this approach, a class-aware band selection and reduction (CBSR) technique is developed for optimal hyperspectral feature representation. A double-branched convolutional Gaussian Bernoulli deep belief network (CGBDBN) is then used for hierarchical spatial feature extraction from LiDAR-derived data and the CBSR data. Using stacked ensemble learning, spatio-spectral features are generated from the two feature streams via a fusion rule and then classified — the results of which are used in labeling a raw 3D LiDAR point cloud through projection. To evaluate this study, extensive experiments were conducted on the IEEE 2018 Houston dataset — the only publicly available dataset with both hyperspectral image (HSI) and 3D point cloud covering the same area — for urban scene classification. The results indicate that the developed CBSR attained comparatively competitive results with state-of-the-art approaches, thus making it a robust spectral feature representation technique. Also, the weight-sharing property, probabilistic modeling, and hierarchical nature of CGBDBN gives our approach the ability to capture high-level contextual features. Furthermore, compared to the spatial- or spectral-only features, the generated spatio-spectral features are more discriminative and significantly aided in improving the proposed model’s efficacy. Overall, the proposed approach, based on the evaluation metrics, is a robust and effective approach for both coarse- and fine-grained raw LiDAR point cloud labeling tasks.

Improving the urban features classification accuracy by fusion of optical and radar high spatial resolution images

Improving the urban features classification accuracy by fusion of optical and radar high spatial resolution images

Inertia and Rationality: Behavioral Confusion and Optimization Paths of Urban Garbage Classification in the Perspective of Accessibility

Garbage classification is in line with China's carbon peaking and carbon neutrality goals and ecological advancement, but during its implementation, it faces the dilemma of a mismatch between supply and demand. Based on the theoretical framework of public service accessibility, this paper uses the scale definition method and SPSS software to empirically study the accessibility of scope, cognitive accessibility, and content of garbage classification in the city of Z as a case study. The results show that the affordability of garbage classification is better than the accessibility of geographic distance and time, and the affordability is significantly and positively correlated with residents' income; the cognitive accessibility is good overall, and the willingness and habit of sorting are average; the accessibility of facilities is better than the accessibility of service items. Given the problems of accessibility of garbage classification in Z city, relevant countermeasures are suggested.

Revealing the Influence of the Fine-Scale Built Environment on Urban Rail Ridership with a Semiparametric GWPR Model

There is a causal interaction between urban rail passenger flow and the station-built environment. Analyzing the implicit relationship can help clarify rail transit operations or improve the land use planning of the station. However, to characterize the built environment around the station area, existing literature generally adopts classification factors in broad categories with strong subjectivity, and the research results are often shown to have case-specific applicability. Taking 154 stations on 8 rail transit lines in Xi’an, China, as an example, this paper uses the data sources of multiple open platforms, such as web map spatial data, mobile phone data, and price data on house purchasing and renting, then combines urban land classification in the China Urban Land Classification and Planning and Construction La1d Standard to classify the land use in the station area using structural hierarchy. On the basis of extracting fine-grained factors of the built environment, a semi-parametric Geographically Weighted Poisson Regression (sGWPR) model is used to analyze the correlation and influence between the variation of passenger flow and environmental factors. The results show that the area of Class II residential land (called R2) is the basis for generating passenger flow demand during morning and evening peak periods; The connection intensity between rail transit station area and bus services has a significant impact on commuters’ utilization level of urban rail transit. Furthermore, two scenarios in practical applications will be provided as guidance according to the research results. This study provides a general analytical framework using urban multi-source data to study the internal relationship and impact between the built environment of urban rail transit stations and passenger flow demand.

Classifying Urban Functional Zones Based on Modeling POIs by Deepwalk

Developing urban functional zone classification method to study urban spatial structure is a hotspot in current research. Using the word embedding model to excavate spatial relationship of the geographic elements in urban functional zones is an important way to develop urban functional zone classification method. However, in these studies, the spatial relationship of geographic elements was regarded as their homogeneity, while the structural similarity of geographical elements was ignored, which inevitably reduces the classification accuracy of urban functional zone classification method. This paper proposes to develop an urban functional zone classification method based on Deepwalk model, which could extract homogeneity and structural similarity of nodes in graph. The proposed method uses POI data to represent geographical elements, organizes POIs into graphs, and uses Deepwalk to embedding POIs for urban functional zone classification. It was applied to classify the urban functional zones of Chaoyang district in Beijing; and the classification results were compared with those of two baseline method based on Word2vec model and Place2vec model. The experimental results show that considering both the homogeneity and structural similarity of geographical elements, the proposed model has higher accuracy than the models only considering the homogeneity of geographical elements.

Effectiveness of landscape indicators for explaining the variability of benthic macro-invertebrates in urban streams

Landscape indicators – measures of land use and land cover – are widely used as proxies for monitoring urban stream conditions, particularly for benthic invertebrates which are often negatively impacted by watershed urbanization. However, multi-scale and nonlinear relationships between benthic macroinvertebrates and landscape configuration derived from fine spatial resolution land cover are not well explored. Here, we developed a series of landscape indicators and assessed their effectiveness in explaining the variability of benthic macroinvertebrate communities in 63 streams across the Greater Vancouver Region in British Columbia, Canada. We asked: 1) How effective are multi-scale landscape indicators in explaining the variability of instream benthic macroinvertebrates? 2) Does the explanatory power of landscape indicators vary at different spatial scales? 3) Do different urban forest classification schemes and their spatial configurations impact the explanatory power of landscape indicators? We developed high spatial resolution (5-m) landscape indicators and evaluated their utility in statistical models explaining taxa richness, instream benthic indices of biological integrity (B-IBI), % Ephemeroptera, Plecoptera, and Trichoptera (EPT), and % Oligochaetes. For all benthic responses, landscape indicators measured at the watershed scale explained 5–25% more variation than riparian-based indicators. Combining indicators mapped at multiple scales further improved the explanatory power of landscape indicators for % EPT and % Oligochaetes, ultimately explaining over 70% of the variability of benthic macro-invertebrates in streams. Distinguishing deciduous and coniferous forest types improved the explanatory power of landscape indicators in a riparian model for B-IBI by 10%. When considering the spatial arrangement of land cover, patch density of forests in the surrounding watersheds of stream explained as much as 47% of the variability in % Oligochaetes. Our results highlight the importance of investigating nonlinear relationships between benthic macroinvertebrates and landscape configuration. This monitoring approach is transferable across cities interested in maintaining the ecological health of urban streams while supporting urban expansion and growth.

Resolution invariant urban scene classification using Multiview learning paradigm

Urban scene classification is an interesting area in computer vision. The task involves classifying a scene from a pair of aerial-view and ground-view images. Existing approaches have considered single-view and Multiview methods. Multiview approaches have shown to be more robust than single-view. However, most existing Multiview approaches neglect the disparity in the resolution of both views. The aerial-view images are captured with sophisticated high-resolution remote sensing devices. While the ground-view images are captured from closer perspectives with lower resolution. This paper proposed a Multiview scene classification (MuSC) model that caters to the resolution disparity in both views. MuSC introduces a Fourier convolution network (FCN) that is robust to variation of resolution in the cross-view images. The FCN is designed to extract local features (in spatial domain) and global features (in spectral domain). The proposed MuSC has a two-stage classifier. The first stage trains a discriminative view-specific network and classifies each view separately. However, the outputs from each view-specific network are projected into a unified subspace and mutual agreement between them is incentivized through contrastive learning. The second stage integrates the predictions from each view-specific network and trains a unified classifier for final prediction. This integration encourages cross-view complementarity. MuSC is evaluated on two datasets, AiRound and CV-BrCT. Several experiments are conducted with different settings. The results demonstrate that MuSC outperforms existing state-of-the-art models.

A novel urban mobility classification approach based on convolutional neural networks and mobility-to-image encoding

Over the last few decades, the classification and prediction of mobility trajectories in dynamic networks have become major research topics. Switching of mobility areas (hand-over) in modern cellular networks is frequent due to restricted coverage area and node speeds (urban, highway, etc.). Accurate management of hand-over events is highly desirable to improve the system’s quality of service. We have exploited the high accuracy of machine learning to classify user mobility from mobility traces which we encoded into images. The method delivers high performance in mobility classification/prediction (exceeding 95%) and avoids the need to study and implement a dedicated neural network structure. The technique requires the conversion of mobility traces into image structures and the subsequent application of a convolutional neural network. We propose a novel approach to classifying mobility that involves data-to-image encoding and machine learning for image classification. Numerous simulations were performed to demonstrate the benefits of the proposed technique and to illustrate the variance in the accuracy of the functions of many encoding/classification parameters. The work represents a first preliminary step towards a new mobility prediction approach. We demonstrate that it is possible to achieve a very high level of prediction accuracy with low computational complexity, exploiting the strength of neural networks in image recognition.

Towards Carbon-Neutral Cities: Urban Classification Based on Physical Environment and Carbon Emission Characteristics

Cities are highly industrialized and populated areas and major sources of greenhouse gas emissions. For carbon neutrality, examining the correlation between urban characteristics and greenhouse gas emissions is necessary. This study aimed to analyze the characteristics of each city from a carbon neutrality perspective. As such, we conducted a carbon-neutral city analysis. First, the physical environmental variables of 250 municipal, county, and district local governments were collected and constructed and then reduced and purified through factor analysis. Second, the type was derived by performing cluster analysis on the reduced factor variables and carbon emissions by analysis unit. Finally, the characteristics of each type were analyzed, and the carbon-neutral city planning and applicable carbon-neutral technology fields were proposed according to the characteristics. After the categorization of carbon-neutral cities throughout Korea, six cluster types were derived; cities in each cluster had similar characteristics. This study suggests that solutions for carbon reduction should be applied by comprehensively considering the social, economic, and environmental characteristics of each city. It concludes that regional physical environmental indicators and energy consumption statistics can be used comprehensively to establish effective policies and apply technologies and techniques at the local government level.

Place-level urban–rural indices for the United States from 1930 to 2018

Rural-urban classifications are essential for analyzing geographic, demographic, environmental, and social processes across the rural–urban continuum. Most existing classifications are, however, only available at relatively aggregated spatial scales, such as at the county scale in the United States. The absence of rurality or urbanness measures at fine spatial resolution poses significant problems when the process of interest is highly localized, as with the incorporation of rural towns and villages into encroaching metropolitan areas. Moreover, existing rural–urban classifications are often inconsistent over time, or require complex, multi-source input data (e.g., remote sensing observations or road network data), thus, impeding the longitudinal analysis of rural–urban dynamics. In order to address this gap, we compare existing rural–urban classifications in the US, and we develop a set of distance- and spatial-network-based methods for consistently estimating the remoteness and rurality of places at fine spatial resolution, over long periods of time, aiming to provide and evaluate temporally consistent rural–urban classifications at fine spatial granularity, but scalable to arbitrary, coarser spatial units. We demonstrate the utility of our approach by constructing indices of urbanness for over 28,000 places in the United States from 1930 to 2018 and further test the plausibility of our results against a variety of evaluation datasets. We call these indices the place-level urban–rural indices (PLURAL) and make the resulting code and datasets publicly available so that other researchers can conduct long-term, fine--grained analyses of urban and rural change. In addition, due to the simplistic nature of the input data, these methods can be generalized to other time periods or regions of the world, particularly to data-scarce environments.

Evaluation of the impact of urban morphology on commercial building carbon emissions at the block scale – A study of commercial buildings in Beijing

With the rapid development of cities, the proportion of building carbon emissions in the total carbon emissions of cities is increasing, and the carbon emission intensity of commercial buildings is higher than that of other building types. The relationship between commercial building life-cycle carbon emissions and urban form factors at the block scale and the spatial differentiation of building carbon emissions in commercial blocks remain unclear. In this study, a calculation method for building life-cycle carbon emissions in urban commercial blocks was established using (1) GIS technology to establish a commercial district building model and urban form factors and (2) block model classification according to architectural layout and combination as well as typical layout pattern extraction through morphological type. Additionally, (3) building life-cycle boundary and scenario parameters were determined, and the urban modeling interface (UMI) platform was used to numerically simulate block samples and typical layout modes for carbon emission calculation model construction. Beijing was chosen as a calculation and analysis example. The results indicated that building shape factor (BSF) is an important indicator of carbon emissions and that the impact of building density (BD) on carbon emissions is two-sided. The spatial differentiation of carbon emissions is obvious, which could provide effective suggestions for target building planning. By studying carbon emission indicators under typical layout modes, we found that low-rise building + tower-enclosed (L + T-E) and low-rise building + tower-determinant (L + T-D) are the best layout modes.

Changes in Pregnancy-Related Mortality Associated With the Coronavirus Disease 2019 (COVID-19) Pandemic in the United States.

To examine pregnancy-related mortality ratios before (January 2019-March 2020) and during (April 2020-December 2020 and 2021) the coronavirus disease 2019 (COVID-19) pandemic overall, by race and ethnicity, and by rural-urban classifications using vital records data. Mortality and natality data (2019-2021) were obtained from the Centers for Disease Control and Prevention's WONDER database to estimate pregnancy-related mortality ratios, which correspond to any death during pregnancy or up to 1 year after the end of a pregnancy from causes related to the pregnancy per 100,000 live births. Pregnancy-related mortality ratios were determined from International Classification of Diseases, Tenth Revision codes A34, O00-O96, and O98-O99. Overall pregnancy-related mortality ratios were partitioned by whether COVID-19 was listed as a contributory cause, and quarterly estimates were compared between 2019 and 2021. Pregnancy-related mortality ratios were compared by race and ethnicity and rural-urban residence before (2019-March 2020) and during (April 2020-December 2020 and 2021) the COVID-19 pandemic. Pregnancy-related mortality was significantly higher in 2021 (45.5/100,000 live births) compared with during the pandemic in 2020 (36.7/100,000 live births) and before the pandemic (29.0/100,000 live births). Pregnancy-related mortality ratios increased across all race and ethnicity and rural-urban residence categories in 2021. The largest increase occurred among American Indian/Alaska Native people during 2021 compared with April-December of 2020 (pregnancy-related mortality ratio 160.8 vs 79.0/100,000 live births, 104% relative change, P =.017). Medium-small metropolitan (52.4 vs 37.7/100,000 live births, 39.0% relative change, P <.001) and rural (56.2 vs 46.5/100,000 live births, 21.0% relative change, P =.05) areas had a larger increase in 2021 compared with April-December 2020 compared with large urban areas (39.1 vs 33.7/100,000 live births, 15.9% relative change, P =.009). Pregnancy-related mortality ratios increased more rapidly in 2021 than in 2020, consistent with rising rates of COVID-19-associated mortality among women of reproductive age. This further exacerbated racial and ethnic disparities, especially among American Indian/Alaska Native birthing people.

Does high-speed railway impact urban logistics industry agglomeration? Empirical evidence from China's prefecture-level cities

Based on the panel data of 280 prefecture-level cities in China from 2007 to 2019, we took the opening of high-speed railways (HSRs) as the “quasi-natural test” and employed propensity score matching (PSM)-difference in difference (DID) method to empirically test the agglomeration effect of the opening of HSRs on the urban logistics industry and the heterogeneity of cities according to location types and urban grade classification. The results show that the opening of HSRs has an impact on the spatial layout of logistics industry by promoting urban economic growth and gathering labor. In the dynamic balance of agglomeration and diffusion, the logistics industry gradually spreads to the surrounding areas of the city through the opening of HSRs. The opening of HSRs has a heterogeneous impact on the agglomeration effect of urban logistics industry. For the rapidly developing first-tier, second-tier and third-tier cities, the opening of HSRs has a significant impact on the agglomeration of their logistics industry and promotes its diffusion to those surrounding cities. On this basis, this paper puts forward corresponding suggestions for the development of the logistics industry and tries to make full use of the development opportunity brought by the opening of HSRs to promote the development of the logistics industry.

Urban Functional Zone Classification Based on POI Data and Machine Learning

The identification of urban spatial functional units is of great significance in urban planning, construction, management, and services. Conventional field surveys are labour-intensive and time-consuming, while the abundant data available via the internet provide a new way to identify urban spatial functions. A major issue is in determining point of interest (POI) weights in urban functional zone identification using POI data. Along these lines, this work proposed a recognition method based on POI data combined with machine learning. First, the relationship between POI data and urban spatial function types was mapped, and the density of each type of POI was calculated. Then, the density values of each type of POI in the study unit were used as feature vectors and combined with the Kstar algorithm to identify urban spatial functions. Finally, the identification results were validated by combining multiple sources of POI data. From the acquired sampling results, it was demonstrated that the proposed method achieved an accuracy of 86.50%. The problem of human bias was also avoided in determining POI weights. High recognition accuracy was achieved, making urban spatial function recognition more accurate and automatable.

Participatory Sensing Based Urban Road Condition Classification using Transfer Learning

Participatory Sensing Based Urban Road Condition Classification using Transfer Learning