International Journal of Product Development (17 papers in press)

Regular Issues

• Review of Models and Frameworks for Set-Based Design  
  by Mayara Silvestre De Oliveira, Fernando Forcellini, Jaime Lozano, Jader Riso Barbosa Jr. 
  Abstract: This paper aims to investigate the state-of-the-art in models and frameworks for set-based design and identify the main gaps and contributions in the literature. As a result, 121 models were analysed. Most models are quantitative, computational, engineering design-oriented, and focus on early stages. Although the narrowing down process plays a central role in the set-based design, very little is addressed regarding its management processes. No model was found describing the inputs and outputs of the set-based design and the narrowing down process simultaneously. Thus, knowledge is dispersed and focused on specific parts. The relevance of this study relies on providing a comprehensive investigation of the state-of-the-art, identifying opportunities to advance in this study field, and providing recommendations for future works seeking to support the development of new methods for implementing and managing set-based design, enabling and encouraging its adoption.
  Keywords: set based design; lean; product development; lean design; literature review; narrowing down process; concurrent engineering; model; framework; strategy.
  
  
• A Formal Consideration of User Tactics During Product Evaluation in Early-Stage Product Development  
  by Trent Owens, Christopher Mattson, Carl Sorensen, Michael Anderson 
  Abstract: Frequent and effective design evaluation is foundational to the success of any product development effort. Products used, installed, or otherwise handled by humans would benefit from an evaluation of the product while formally considering both the physical embodiment of the technology, termed technology, and the steps a user should take to use that technology, termed tactics. Formal and simultaneous evaluations of both technology and tactics are not widespread in the product design literature. Although informal evaluation methods have advantages, formal methods are also known to be effective. In this paper we propose a formal method for evaluating tactics and technology simultaneously. Unlike the published literature, this evaluation involves explicitly defined tactics in the form of a written description of the actor, environment, and series of steps. It also involves the use of stage-appropriate, explicitly defined tactics-dependent criteria, which include criteria from a broad range of impact categories, such as impacts on the user, environment, project, and technology.
  Keywords: Conceptual Design Evaluation; Tactics Evaluation; Human-Centered Design; Human Factors.
  
  
• Study on evaluation method of human-computer interface quality of intelligent products based on Bayesian classification  
  by Jinrong Li  
  Abstract: Starting from effectively improving the evaluation accuracy and efficiency of the quality of the interactive interface of intelligent products, this paper proposes a method for evaluating the quality of the interactive interface of intelligent products based on Bayesian classification. An adaptive Gaussian filter is introduced to adjust the color of the interactive interface of intelligent products through logarithmic operators, and the interactive interface of intelligent products is formally described. The Bayesian classification method is used to build an evaluation model for the quality of the interactive interface of intelligent products, and the quality of the interactive interface of intelligent products is evaluated according to the Bayesian classification probabilistic inference mechanism. The experimental results show that the mean saliency of the proposed method is as high as 95.7%, the recognition accuracy rate reaches 96.4%, and the evaluation time is only 7.2s, which has a good evaluation effect.
  Keywords: Bayesian classification; adaptive Gaussian filter; intelligent product; human-computer interaction interface; quality assessment.
  DOI: 10.1504/IJPD.2023.10058408
   
  
• Prediction method of product market demand based on Prophet random forest  
  by Chaoyong Jia  
  Abstract: This paper proposed a prediction method of product market demand based on Prophet random forest. On the basis of analyzing the workflow of Prophet model and the forest generation and decision-making process of random forest algorithm, the raw data of product market are preprocessed through the process of data filling, feature standardization and feature mapping. Then, the optimal subset selection algorithm is used to extract the product market demand characteristics, and the demand characteristics are input into Prophet random forest to realize the prediction of product market demand. The experimental results show that the prediction accuracy of the proposed method is 0.969, and the maximum prediction time in the experiment is only 14.9s, and the waveform trend of the predicted result is roughly the same as that of the actual value, which highlights the effectiveness of the proposed method.
  Keywords: Prophet model; Random forest algorithm; Optimal subset selection algorithm; Product market demand; Demand forecast.
  DOI: 10.1504/IJPD.2023.10058409
   
  
• Fuzzy edge detection method of product packaging image based on Kalman filter  
  by Ruifang Xing, Jingjing Feng, Yayun Fan 
  Abstract: In order to improve the quality, effect and efficiency of fuzzy edge detection of product packaging image, a fuzzy edge detection method of product packaging image based on Kalman filter is designed. Firstly, singular value decomposition algorithm is used to remove the noise of product packaging image. Then the Fourier spectrum of the product packaging image is obtained by FFT operation, and the image blur parameters are quickly identified. Finally, the image fuzzy edge is processed by Kalman filter to realize image fuzzy edge detection. The experimental results show that the detection signal-to-noise ratio of the proposed method is as high as 61.5db, the quality factor is as high as 0.97, and the detection time is short, only 19.7s. It can be proved that the proposed method can effectively improve the quality and efficiency of fuzzy edge detection of product packaging image.
  Keywords: Kalman filter; singular value decomposition; FFT; product packaging image; fuzzy edge detection.
  DOI: 10.1504/IJPD.2023.10058410
   
  
• Color matching method of product interactive interface based on user experience  
  by Yun Xu 
  Abstract: In this paper, a color matching method of product interactive interface based on user experience is proposed. Firstly, the color matching visual information of product interaction interface is collected in the adjacent block area. According to the visual information collection results, the color matching element features of product interaction interface are extracted by wavelet transform. Secondly, the RGB feature decomposition model of color matching elements in product interaction interface is established to complete the visual information fusion of color matching in product interaction interface; Finally, considering the user experience, the color matching reconstruction model of product interactive interface is constructed to realize the color matching of product interactive interface. The experimental results show that the proposed method has high color gamut coverage and image signal-to-noise ratio, high user experience satisfaction, and good color matching quality and effect of product interaction interface.
  Keywords: User experience; Visual information fusion; Wavelet transform; Product interaction interface; Color matching.
  DOI: 10.1504/IJPD.2023.10058412
   
  
• The deep Mining of Consumer Behavior Data on Product Network Marketing Platform  
  by Chunming Yu, Xin Jin 
  Abstract: To overcome the problems of low accuracy of behavior analysis results and low purchase proportion in traditional methods, a deep mining method of consumer behavior data on product network marketing platform is proposed. Firstly, according to the nearest neighbor data distribution, the relevant subspace of consumer behavior data of product e-marketing platform is divided, and the sparsity difference of relevant subspace data in each attribute is calculated. Then, the filtering of outlier interference data is completed by setting the difference threshold of local sparsity factor. Finally, the multi-source data mining method is used to analyze the difference between the data attribute weight and the target weight of each attribute behavior, so as to realize the in-depth mining of consumer behavior data. Test results show that the maximum error of consumer behavior analysis of the design method is only 1, and the purchase proportion after secondary marketing reaches 10.2%
  Keywords: Network marketing platform; Consumer behavior data; Deep mining; Nearest neighbor data; Local sparse factor; Outlier data.
  DOI: 10.1504/IJPD.2023.10058413
   
  

Special Issue on: AI/Machine Learning Techniques as a Tool for Process Modelling and Product Design

• Color offset compensation method of product packaging image based on color difference interpolation  
  by Zhao Lan 
  Abstract: In order to effectively improve the accuracy of image color offset compensation and reduce the compensation time, a color offset compensation method based on color difference interpolation is proposed in this paper. Based on bilinear chromatic aberration interpolation algorithm, the full resolution chromatic aberration signal of image is estimated preliminarily. After image filtering, the basic color data and gray balance data of the image are compared to obtain the color offset. Taking color block as the minimum unit of deviation compensation, the average value of extended color level interpolation was added to the single color level of the calibrated image to complete color offset compensation. Experimental results show that the maximum compensation time of the proposed method is only 29.6s, and the maximum MSE value of the compensation result is only 0.179, indicating that the proposed method has higher compensation accuracy.
  Keywords: Chromatic interpolation algorithm; Linear filtering; Product packaging image; Color offset compensation.
  
  

Special Issue on: Smart Products

• An Interactive Method of Digital Fitness Equipment Based on Machine Learning  
  by Rui Zhang, Fang Liu, Cheng Wang 
  Abstract: Aiming at the problems of long operation time, high operation error rate and low user satisfaction in traditional methods, this paper proposes an interactive method of digital fitness equipment based on machine learning. Firstly, the paper designs an interactive platform for digital fitness equipment to collect user data. Then, the consistent convergence model is used to adjust the user's human-computer interaction requirements, the machine learning algorithm is used to design the digital fitness equipment, and the adaptive controller is used to control the user's human-computer interaction behavior. Finally, the human-computer interaction interface of digital fitness equipment is designed. The experimental results prove that the interactive interface operation time measured by the design method is short, the shortest time is only 2.51s; The error rate range of interactive interface operation is 0-0.7%, and user satisfaction is high, indicating that this method has a better application effect.
  Keywords: machine learning; human-computer interaction; digital fitness equipment; consistent convergence; adaptive controller.
  
  
• The 3D Animation Display of Smart Products Based on Hanoi Tower Recursive Algorithm  
  by Qiong Wang 
  Abstract: In the process of using traditional methods to display smart product 3D animation, the image is not preprocessed, resulting in poor animation display effect and low efficiency. Therefore, a smart product 3D animation display method based on the Hanoi Tower recursive algorithm is proposed. Using the weighted average method to grayscale the 3D image of the smart product, obtain the grayscale image, and combine the Harris corner detection method to extract the image feature points, and reconstruct the 3D image of the smart product based on the feature extraction results. The Hanoi Tower recursive algorithm is used to construct a 3D animation display model to realize the 3D animation display of smart products. The simulation results show that the 3D animation display image of the smart product of this method has higher clarity, higher display effect and higher display efficiency.
  Keywords: Hanoi Tower; Recursive algorithm; Smart products; 3D animation display; Graying processing; Harris corner detection.
  
  
• A fault diagnosis method of intelligent electronic equipment based on dynamic fusion  
  by Jun Qian, Kaijun Nie 
  Abstract: In order to overcome the problems of low accuracy of fault data extraction and long diagnosis time, a fault diagnosis method of intelligent electronic equipment based on dynamic fusion is proposed. Convolution neural network and long-term and short-term memory neural network are dynamically fused to extract the characteristic parameters of intelligent electronic equipment based on neural network. Combined with the principle of feature collection, the fault information of intelligent electronic equipment is collected, the covariance matrix of data noise data is constructed, the key noise is extracted, the noise reduction of fault data is completed, and the fault diagnosis of intelligent electronic equipment is carried out according to the results of fault information fusion and fuzzy kernel c-means clustering. The experimental results show that the highest accuracy of fault data extraction is about 95%, and the shortest time of sample fault data extraction is about 0.5s.
  Keywords: Dynamic fusion; Intelligent electronic equipment; Noise interference degree; Fuzzy kernel c-means clustering; fault diagnosis.
  
  
• Appearance design defect detection of rigid plastic products based on geometric feature extraction  
  by Jiang-lin Hua, Yuan-sheng Liu 
  Abstract: In order to overcome the problem of low geometric feature extraction accuracy and defect detection accuracy of traditional appearance design defect detection methods, a new appearance design defect detection method of rigid plastic products based on geometric feature extraction is proposed in this paper. This method first analyzes the types of design defects of rigid plastic products, and then extracts the defect area, the length and width of the defect area, the gradient characteristics and shape characteristics of the defect area. Based on the geometric feature extraction results, combined with Gaussian Laplace algorithm, the appearance design defects of rigid plastic products are detected. The experimental results show that compared with the traditional design defect detection methods, this method can accurately extract the geometric features of defects, so as to obtain more accurate defect detection results, and the maximum defect detection accuracy can reach 97%.
  Keywords: Geometric feature extraction; Hard plastic products; Appearance design; defect detection.
  
  
• Expression capture of virtual reality equipment based on expression weighted distance SLLE  
  by Shuitian Yan 
  Abstract: In order to overcome the problems of long time-consuming and high error in traditional expression capture methods, an expression capture method for virtual reality devices based on expression weighted distance SLLE is proposed. The LM algorithm is used to generate the virtual viewpoint of facial expression, and the adaptive weighted LDP method is used to process the facial expression features, remove the redundant information and retain the key features of the image. Considering the differences between different image types and viewpoints, the weighted distance SLLE method is used to capture facial expression images. Experimental results show that this method can effectively capture the viewpoint in facial expression, and the time is short and the expression capture error is small.
  Keywords: expression weighted distance SLLE; virtual reality equipment; adaptive weighting; face viewpoint.
  
  
• Double fingerprint accurate identification method of intelligent door lock based on spatio-temporal information feature fusion  
  by Liuping Cai, Fenghui Huang 
  Abstract: Aiming at the problems of low recognition accuracy and large feature extraction error in the dual fingerprint recognition of intelligent door lock, a dual fingerprint accurate recognition method of intelligent door lock based on spatio-temporal information feature fusion is proposed. Set the standardized form of dual fingerprint image, calculate the gray mean and variance of fingerprint image, and complete fingerprint image preprocessing; According to the coordinate relationship of the dual fingerprint direction field of the intelligent door lock, the dual fingerprint block of the intelligent door lock is divided, and the feature extraction is completed by solving the gradient value of the dual fingerprint block. The temporal and spatial features of fingerprint images are fused, and the fusion results are corrected with the help of calibration function to complete accurate fingerprint identification. The results show that the recognition accuracy of the proposed method is good.
  Keywords: spatiotemporal information feature fusion; Intelligent door lock; Double fingerprint; Equalization; Calibration function.
  
  
• A real-time acquisition method of local change features of 3D animation considering product design  
  by Weijia Chai, Zhiqiang Yue 
  Abstract: In order to improve the poor matching effect of traditional feature acquisition methods, a real-time feature acquisition method of 3d animation local changes considering product design is proposed in this paper. Firstly, the local variation characteristics of the animation are described according to the product design concept. Then, the information transferred in the design of the product is analyzed, and the anti-noise ability of the change feature descriptor is fitted by gaussian difference function. Finally, the expression factor of product appearance is transformed into the extreme point information of the change feature, and the extreme point equation of the local change feature is obtained. The real-time collection of local variation features is realized by considering product design rules. The results show that the matching effect of the proposed method is 200 dimensions higher than that of traditional methods, and the recognition rate is 40%~10% higher than that of traditional methods.
  Keywords: product design; 3D animation; Local variation characteristics; Real-time acquisition; Data collection; Sampling method.
  
  
• Research on secure storage data access control of smart TV based on blockchain  
  by Jianxin Qiu 
  Abstract: In order to solve the problems of low data security, access control accuracy and long data query time existing in traditional data access control methods, this paper proposes a secure storage data access control of smart TV based on blockchain. Firstly, the naive Bayesian algorithm is used to calculate the posterior probability according to the data safety level; Secondly, the category corresponding to the maximum posterior probability is regarded as the category of the data, and the data classification is realized by iterative calculation; Finally, on the basis of data classification, different categories of smart TV secure storage data access control are realized through ciphertext encryption, data decryption, access policy update and other steps based on blockchain. Experimental results show that the method has high data security, the average accuracy of data access control is 96.3%, and the data query time is always less than 0.5s.
  Keywords: Blockchain; Smart TV; Secure storage of data; Access control; Naive Bayesian algorithm; Ciphertext encryption; Data decryption.
  
  
• A Color Enhancement Method of Plane Image Based on the Interactive Interface of Mobile Smart Devices  
  by Jianjun Wang 
  Abstract: When using the existing methods to enhance the image color of the interactive interface of mobile intelligent devices, there are some problems, such as long enhancement time, easy to produce image distortion and so on. Therefore, a color enhancement method for planar image of interactive interface of mobile intelligent device is proposed. The target image is collected by image sensor, the plane image is transformed from high-dimensional subspace to low-dimensional subspace, the linear map is established, the projection distance between each single pixel in the image to be processed is obtained, and the image is processed by local dimension reduction method to complete the preprocessing; The histogram method is used to adaptively adjust the dark area in the interface, and the enhancement is completed by mapping function and median filter method. The results show that this method has low image distortion rate and short time in image enhancement processing.
  Keywords: mobile smart device; interactive interface; plane image; color enhancement; image acquisition.