Training Computers to See the Built Environment Related to Physical Activity: Detection of Microscale Walkability Features Using Computer Vision.
Training Computers to See the Built Environment Related to Physical Activity: Detection of Microscale Walkability Features Using Computer Vision.
复制标题
训练计算机识别与身体活动相关的建成环境：利用计算机视觉检测微观步行适宜性特征
DOI：
10.3390/ijerph19084548
复制
发表时间：
2022-04-09
影响因子：
--
通讯作者：
Middel, Ariane
中科院分区：
文献类型：
--
作者：
Adams, Marc A.;Phillips, Christine B.;Patel, Akshar;Middel, Ariane
研究目的是训练和验证一种深度学习方法，以检测与行人身体活动相关的微尺度街景特征。这项工作通过将计算机视觉技术与谷歌街景（GSV）图像相结合来克服进行审计的障碍（例如，时间，安全性和专家劳动力成本）。在microscale Audit of Pedestrian streetscape Mini工具的指导下，使用EfficientNETB5架构为8个微尺度特征构建深度学习模型：人行道、人行道缓冲、路缘切割、斑马线和人行横道、步行信号、自行车标志和路灯。我们使用了训练校正循环，即在训练数据集上训练图像，使用单独的验证数据集进行评估，并进一步训练，直到达到可接受的性能指标。此外，我们使用训练过的模型来审计WalkIT亚利桑那州试验中的参与者（N = 512）社区。探讨了微尺度特征与gis测量和参与者报告的社区宏观步行性之间的相关性。分类器的准确率、召回率和总体准确率均超过bb0.84%。总体微观尺度与总体宏观尺度步行适宜性相关（r = 0.30, p < 0.001）。模型检测和自我报告的人行道（r = 0.41, p < 0.001）与人行道缓冲（r = 0.26, p < 0.001）呈正相关。计算机视觉模型的结果提出了一种替代训练有素的人类评估师的方法，允许对数百或数千个社区进行人口监测或假设检验的审计。
The study purpose was to train and validate a deep learning approach to detect microscale streetscape features related to pedestrian physical activity. This work innovates by combining computer vision techniques with Google Street View (GSV) images to overcome impediments to conducting audits (e.g., time, safety, and expert labor cost). The EfficientNETB5 architecture was used to build deep learning models for eight microscale features guided by the Microscale Audit of Pedestrian Streetscapes Mini tool: sidewalks, sidewalk buffers, curb cuts, zebra and line crosswalks, walk signals, bike symbols, and streetlights. We used a train–correct loop, whereby images were trained on a training dataset, evaluated using a separate validation dataset, and trained further until acceptable performance metrics were achieved. Further, we used trained models to audit participant (N = 512) neighborhoods in the WalkIT Arizona trial. Correlations were explored between microscale features and GIS-measured and participant-reported neighborhood macroscale walkability. Classifier precision, recall, and overall accuracy were all over >84%. Total microscale was associated with overall macroscale walkability (r = 0.30, p < 0.001). Positive associations were found between model-detected and self-reported sidewalks (r = 0.41, p < 0.001) and sidewalk buffers (r = 0.26, p < 0.001). The computer vision model results suggest an alternative to trained human raters, allowing for audits of hundreds or thousands of neighborhoods for population surveillance or hypothesis testing.
登录
查看更多内容
影响因子：
2.1
作者：
Alfonzo, Mariela;Boarnet, Marlon G.;Anderson, Craig L.
通讯作者：
Anderson, Craig L.
影响因子：
5.4
作者：
Cain, Kelli L.;Millstein, Rachel A.;Sallis, James F.;Conway, Terry L.;Gavand, Kavita A.;Frank, Lawrence D.;Saelens, Brian E.;Geremia, Carrie M.;Chapman, James;Adams, Marc A.;Glanz, Karen;King, Abby C.
通讯作者：
King, Abby C.
DOI：
10.1186/1479-5868-6-32
发表时间：
2009-06-09
影响因子：
8.7
作者：
Cerin, Ester;Conway, Terry L.;Sallis, James F.
通讯作者：
Sallis, James F.
影响因子：
2.5
作者：
Berriel, Rodrigo F.;Rossi, Franco Schmidt;Oliveira-Santos, Thiago
通讯作者：
Oliveira-Santos, Thiago
DOI：
10.1007/s11524-010-9505-x
发表时间：
2010-12-01
影响因子：
6.6
作者：
Badland, Hannah M.;Opit, Simon;Mavoa, Suzanne
通讯作者：
Mavoa, Suzanne