Experimental materials
In this study, six animated images from the children’s animated images of “PAW Patrol” are selected as the material for color extraction. Those images were randomly selected in the form of animated screenshots from the main episodes of episodes 1, 17, 33, 58, 76, and 99, which are shown in Figs. 1(a), (b), (c), (d), (e), and (f), respectively. As Children’s favorite entertainment, animation play an important role in the process of children’s growth12. “PAW Patrol” is an action-adventure cartoon series launched by Nickelodeon, which tells the story of a 10-year-old boy, Ryder, who sets up a rescue team with his adopted puppies to help people. The captain in “PAW Patrol” takes on the leader responsibility and has an almost perfectly positive persona, acting as a moral benchmark as well as a code of conduct. The members in “PAW Patrol” become favorite images that children like in hearts with the teamwork spirit of conveying love, responsibility and mutual help. Particularly, the colors of the animated images of “PAW Patrol” leave a deep impression on children27.
In this paper, the children’s shopping cart model with three main colors is selected as the color design product, whose 3D model is shown in Fig. 7. The reasons are as follows. On one hand, more than 20,000 children are injured by adult shopping carts and sent to the emergency room each year in the United States, with most of the injuries occurring in the face and neck. Therefore, professionally designed shopping cart products for children are valued by designers (Supermarket shopping cart).28 On the other hand, the users who watch the animation of “PAW Patrol” match the age of users of children’s shopping cart products. From the emotional needs of children while watching animation, mapping the “PAW Patrol” animation colors to the color scheme of children’s shopping cart products helps to improve children’s product experience and shopping experience29. In addition, the three-color children’s shopping cart can not only attract children’s attention, but also arouse children’s happy feeling when three colors are carefully combined 2, 13.
Methodology
The study was validated and approved by The Biomedical Ethics Committee of Anhui University (BECAHU-2022-003), and all studies were conducted in accordance with the Declaration of Helsinki and other relevant guidelines, and all experimental subjects voluntarily signed the informed consent form after being fully informed before conducting the experiments and subjective studies. All subjects and/or their legal guardians have informed and consented to the publication of identifying information/images in online open access publications.
• Eye-tracking technology
The eye-tracking technology is based on infrared devices and image acquisition devices, which tracks subtle changes in eye features in real time by actively projecting light beams such as infrared rays to the iris. The eye-tracking device used in this study was the Tobii Pro Glasses 2, manufactured in Sweden. The eye-tracking technique has less interferes with the acquisition of the subject’s evaluation process and meanwhile enables the recording of various data during eye movements30.
There are four common methods for data visualization of eye-tracking: AOI method, scan path method, 3D spatial method and thermal zone map method. The hot zone map method is selected for color extraction of animated images. The eye-tracking data visualization method should be reasonably determined according to the actual design evaluation content30–34. In the heat zone map method, the denser the hot spots, the longer the subject’s gaze time in the area of interest, the longer the gaze time, and the greater the interest in the color matching of the area35. Therefore, extracting colors and obtaining color schemes by the heat zone map method of eye-tracking experiments is in line with the expected goal of this study14.
• Multilayer perceptron neural network
Multilayer perceptron neural network (MLP) can be regarded as a logistic regression classifier, which consists of three parts: the input layer, the hidden layer, and the output layer. The input layer receives the external data and calculates the excitation values through the activation function, and then the values are passed to the hidden layer. The hidden layer takes the results from the upper layer as input and calculates the excitation values through the activation function, and the obtained data are passed to the output layer. The nodes of the hidden layer and the output layer can be perceptrons. The multilayer perceptron with more nodes has multiple perceptrons and the output layer can also be adjusted according to the actual application36.
The output layer in MLP often uses the soft max function in many color combination screening or classification tasks. The mathematical expression of MLP is shown in Eq. (2):
$$\left\{ \begin{gathered} {h_j}={\varphi ^{(1)}}(\sum\limits_{{i=1}}^{n} {{x_i} \cdot \omega _{{ji}}^{{(1)}}} +{b^{(1)}}) \hfill \\ {y_k}={\varphi ^{(2)}}(\sum\limits_{{j=1}}^{n} {{h_j} \cdot \omega _{{kj}}^{{(2)}}} +{b^{(2)}}) \hfill \\ \end{gathered} \right.$$
2
where m, n and k represent the numbers of neurons in the input layer, the hidden layer and output layer, respectively. \({h_j}\) is the output value of the hidden layer,\({\varphi ^{(1)}}\) refers to the activation function from the input layer to the hidden layer,\({x_i}\) denotes the external input data,\(\omega _{{ji}}^{{(1)}}\) is the connection weight from neurons in the input layer to neurons in the hidden layer,\({b^{(1)}}\) is the bias value of the hidden layer,\({y_k}\) is the output value of the output layer,\({\varphi ^{(2)}}\) is the activation function from the hidden layer to the output layer,\(\omega _{{kj}}^{{(2)}}\) is the connection weight from neurons in the hidden layer to neurons in the output layer, \({b^{(2)}}\) is the activation function from the hidden layer to the output layer37.
This work adopts an MLP to deal with the mapping relationship problem between user ratings and color combinations for the preparation of multiple color combination selection. The relationship between dependent variables and independent variables in the product color decision is not linear. The MLP is the core of machine learning and deep learning, which has the advantages of self-adaptability, self-learning, real-time, high robustness and so on. Multiple hidden layers facilitate scholars to build powerful models for efficiently solving nonlinear system problems21. All the above studies used MIPs for screening and decision making of practical problems, and the experimental findings demonstrated that the MIP has substantial improvement in stability and accuracy compared with other algorithms37–39. Therefore, screening multiple color combinations for children’s shopping carts by MIPs is in line with the expected goal of this study.