Fig 1.
Framework of the Urban Street Child-Friendliness Assessment System.
Image source: self-drawn by the author.
Fig 2.
Schematic distribution map of urban street research samples in Shanghai urban area.
Image source: self-drawn by the author. This diagram is for illustrative purposes only and may differ from the original image.
Table 1.
Statistical table of child-friendly concrete features in urban streets.
Fig 3.
Example of street view photos of urban street taken on site.
Image source: Taken by the author on site.
Table 2.
Statistical table of the number of urban street samples with different child-friendliness ratings in the urban street child-friendliness assessment model dataset (before data augmentation).
Table 3.
Statistical table of the number of urban street-view samples with different child-friendliness ratings in the pre-trained CNN model dataset (before data augmentation).
Fig 4.
Urban streetscape photos after image augmentation in the pre-trained convolutional neural network model dataset.
Image source: Image transformation performed by computer.
Table 4.
Statistical table of the number of urban streetscape samples with different child-friendliness ratings in the pre-trained convolutional neural network model dataset (after data augmentation).
Fig 5.
Urban streetscape photos after image augmentation in the urban street child-friendliness assessment model dataset.
Image source: Image transformation performed by computer.
Table 5.
Statistical table of urban street samples with different child-friendliness rating levels in the urban street child-friendliness assessment model dataset (after data augmentation).
Fig 6.
Experimental data plot of pre-trained convolutional neural network models on training and validation sets.
Image source: self-drawn by the author.
Table 6.
Experimental data table of pre-trained convolutional neural network model (best performing model on validation sets) on test set.
Fig 7.
Experimental Data Plot of the urban street child-friendliness assessment model on Training and Validation Sets.
Image source: self-drawn by the author.
Table 7.
Experimental data table of the urban street child-friendliness assessment model (best performing model on validation sets) on test set.
Fig 8.
Heat maps of selected urban street scenes in the pre-trained convolutional neural network model.
Image source: self-drawn by the author.
Fig 9.
Schematic diagram of the contribution of concrete features in the urban street child-friendliness evaluation model (Top 20 concrete features with the highest contributions).
Image source: self-drawn by the author.
Fig 10.
Schematic diagram of urban streets in the ancient city area of Suzhou.
Image source: Taken by the author on site.
Table 8.
Sample size statistics for different levels of child-friendliness ratings in the Suzhou Ancient City dataset (before data augmentation).
Table 9.
Sample Size Statistics for Different Levels of Child-Friendliness Ratings in the Suzhou Ancient City dataset (after data augmentation).
Fig 11.
Experimental Data of the Transfer Model for Suzhou Ancient City on the Training Set.
Image source: self-drawn by the author.
Table 10.
Experimental data of the suzhou ancient city transfer model (best-performing model on the training set) on the test set.
Fig 12.
Road numbering map of Kongjiang Road Subdistrict, Yangpu District, Shanghai.
Image source: self-drawn by the author. This diagram is for illustrative purposes only and may differ from the original image.).
Table 11.
Child-friendliness evaluation matrix of urban streets in kongjiang road subdistrict, Yangpu District, Shanghai.
Fig 13.
Street view images of low-rated roads in Kongjiang Road Subdistrict as predicted by the automated evaluation model.
Image source: Taken by the author on site.
Fig 14.
Schematic Distribution Map of Urban Street Samples in Shanghai Used for Comparative Evaluation.
Image source: self-drawn by the author. This diagram is for illustrative purposes only and may differ from the original image.).
Table 12.
Urban street feature statistics for traditional evaluation methods.
Table 13.
Comparison of credibility of evaluation results between the two different evaluation methods.