Computer-aided design of hand-drawn art food packaging design based on a deep neural network model | International Journal for Simulation and Multidisciplinary Design Optimization (IJSMDO)

Open Access

Table 1

Description of related works (source: author).

Ref	Summary	Findings	Limitation
[9]	The study examined the precision of categorizing hand-drawn drawings by applying "Support Vector Machines (SVM)" and "Artificial Neural Networks (ANN)" while utilizing image characteristic extracting approaches.	The research demonstrated findings that were equivalent to current innovative approaches for constructed specifications, achieving a classifications accuracy combining SVM and ANN. The comparison result of Precision (61%), Recall (59%), and F1-Score (59%).	Due to the lack of image characteristic extracting methods, which had been efficient in several image classification problems.
[10]	The research evaluated the issue of recognizing hand-drawn diagrams without an internet connection.	According to the results, DiagramNet outperformed conventional offline approaches that utilized compact images extracted from demanded web databases.	Although the model had been fine-tuned for use with hand-drawn commercial process diagrams, it might not be applicable to other kinds of diagrams.
[11]	The article proposed the laborious and error-prone task of manually inputting hand-drawn electrical and electronic circuit information into machines.	The testing results demonstrated that the suggested technique effectively identified electrical and electronic circuit elements generated manually with an average reliability of 93.83% on the collection.	A notable drawback was that the assessment was dependent on a single dataset, which could not capture all the possible differences in constructed circuit elements.
[12]	The paper presented the approach of a new algorithm dubbed the “Counting Key-Points algorithm (CKP)”.	To demonstrate CKP's authority, its classification efficiency was contrasted with other innovative approaches that depended on hand-drawn images. The comparison result of Accuracy (65%).	The drawback highlighted the need for more exploration and improved statistics to guarantee the CKP algorithm's resilience and generalization.
[13]	The study presented that machine learning (ML) methods were significantly influenced by the setting of” hyperparameters (HPs)".	To demonstrate the advantages of utilizing a large hand-drawn sketching database for classification tasks, their publication gives a quick introduction to significant observation and setup methodologies.	They investigated how the pre-trained graphic geometry group performed when various learning settings, player constraints, “batch normalization (BN)”, and dropouts were used (VGG-19)
[14]	The research highlighted the need to thoroughly enhance the safeguarding of "intellectual property rights (IPRs)".	Findings considered that a new developmental paradigm could only be fostered and society's creativity sparked by a more robust system of intellectual property rights protections.	A drawback of the section was that it only discusses one part of IP rights—specifically, patent and trademark rights in computer nonlinear forecasting methods used for package design.
[15]	The article presented the RANSAC method classification and the enhanced ICP algorithm.	Based on the outcome, it could be inferred that 3D digital technologies had a substantial influence on contemporary package design, resulting in enhanced visual expressiveness and visual effect.	Furthermore, not all possible issues or factors are connected to the use of 3D digital technology in package design.
[16]	The study examined how brand perception was affected by the level of complexity or simplicity of packaging design.	The findings demonstrated that the level of simplicity or complexity in a package design has a substantial influence on how a brand is perceived.	The unique circumstances of the Champagne bottle designs and the investigated attributes might limit the results’ generalizability.
[17]	The research presented a “computer-aided design (CAD)” system specifically designed for graphic art items.	The research contributed to the development of CAD technology for graphic applications design that makes use of AutoCAD and VBA.	To start, the CAD system might not be suitable for individuals without experience with AutoCAD and VBA, as these were the primary methods for improvement.
[18]	The article examined the use of “Extended Reality (XR)” technologies, notably “Virtual Reality” (VR) and “Augmented Reality (AR)”, in commercial environments.	The study's findings pointed to the many advantages of using projection-based AR technologies in collaborative design processes.	Another possible downside of co-creative design practices that might restrict innovation was the difficulty in communicating between designers and non-designers.
[19]	The paper examined several “corporate social responsibility (CSR)” reports to determine if the non-financial reporting of cosmetic “multinational corporations (MNCs)” was centered on the idea of a circular economy and whether CSR filings guaranteed sufficient transparency about circular strategies.	The investigation indicated that multinational corporations pay close attention to circularity when writing their corporate social responsibility reports, which might often outline goals and activities that cover numerous dimensions.	One additional potential drawback was that were only examined a small portion of these companies’ circular economy initiatives in their CSR reports.
[20]	The study evaluated the colors of relaxation food packaging, taking into account both visual and psychological factors.	There was an underlying relationship between the emotional semantic perception of colors and color fulfillment, according to quantitative investigation. The correlation coefficient and significance for V2 and V11, respectively, are 89.6% and 0.006, and 81.6% and 0.025, as the comparison result.	Although the material provided insightful information on how customers perceived the color of snack food packaging, it was important to note that it contains several constraints.


[21]	The study aimed to optimize industrial product decisions using a deep CNN with RBF neural network-based objective weight assignment, converting multi-objective optimization into a single-objective job and producing non-inferior solutions for fine-tuning.	As a result, the algorithm's effectiveness in directing different intelligent system pipelines has been shown, with possible applications such as 3D reconstruction and system optimization.	Implementing deep neural networks in real time and scaling to bigger datasets can be difficult due to their complexity and processing demands.
[22]	The study aimed to enhance agricultural product packaging using computer-aided design, focusing on ecological ideas and local value, using color design theory, intelligent three-dimensional simulation techniques, and a comprehensive package design framework for consumer groups.	Using computer-aided techniques, a model demonstrating improved agricultural product packaging was successfully developed. It incorporates ecological ideas, effective value transmission, and improved visual appearance.	Adopting the computer-aided package design paradigm may present practical implementation hurdles as well as difficulties in adequately capturing a variety of local and cultural features.
[23]	The project aims to improve package design efficiency for marine diesel engines by integrating CADDCAM to address backward NC programming and data communication issues, streamlining the design process and addressing software interoperability and aircraft label occurrence issues.	As a result of the suggested CADDCAM integration solution, marine diesel engine packaging has improved overall design effectiveness, decreased repetitive labour, and increased design efficiency.	The study's scope is restricted to certain problems in the field of marine diesel engine packaging design within the framework of CADDCAM integration.
[24]	The study explores the use of computer-aided design (CAD) techniques for designing goods, focusing on expediting design procedures while maintaining experimental validation. It discusses major property estimating techniques, their relationship to requirements, and case examples.	The outcome demonstrates the effectiveness of the CAD framework in accelerating processes and incorporating experimental findings into product design by demonstrating its practical use.	Despite their early development, the industry does not completely utilize CAD technologies, and their promise is not fully realized.
[25]	The study investigates the integration of marketing and design concepts in the design of sporty sunglasses and wristwatches, utilizing research tools, marketing concepts, and product development processes like mind maps, mood boards, sketching, and 3D CAD modelling.	As a result, the project effectively combines theoretical and practical approaches, exhibiting a thorough design process that converts unprocessed client input into 3D CAD-modeled items and making a contribution to the area of product development.	The specificity of the product emphasis (sporty and classic sunglasses with hand-watches) and the subjective nature of design choices may pose constraints to the generalizability of the research.

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.