An In-Depth Evaluation of Virtual Try-On Systems for E-Commerce: Analyzing Performance, Practicality, and Cost-Effectiveness

Purpose Virtual try-on (VTO) technology offers consumers a shopping experience comparable to direct product examination by providing detailed product information and enhancing enjoyment during online shopping. Against this backdrop, this study extends the technology acceptance model (TAM) to investigate the factors influencing consumers’ attitudes toward VTO technology and their willingness to purchase online. Design/methodology/approach Data were collected through an online survey of 228 Italian respondents. The proposed research model was tested through an exploratory factor analysis (EFA) and a confirmatory factor analysis (CFA), followed by a structural equation model (SEM) with an ordered Probit approach. Findings The results highlight the significant impact of perceived enjoyment (PE), innovativeness and perceived environmental benefits (PEBs) on consumer attitudes toward VTO and their online purchase intentions. This study underscores the role of VTO in enhancing online shopping experiences, leveraging both utilitarian and hedonic values, ultimately encouraging technology adoption. Research limitations/implications The findings offer valuable insights for retailers seeking to encourage online shopping using VTO technology, fostering PE, stimulating interest in innovative solutions and promoting responsible consumption choices. Originality/value This research extends the TAM by integrating external variables – innovativeness and PEBs – into the analysis, while accounting for the mediating role of PE on consumer behaviour.

The proposed virtual try-on model works on synthesizing naturalistic images by superimposition of target clothing item on a person. The proposed model offers a significant advancement in virtual try-on technology, by introducing a lightweight architecture that reduces the system requirements required to run the application. Many different preprocessing methods have been utilized to account for different poses and body shapes. Great efficiency has been achieved by employing the Pix2Pix generative adversarial networks (GAN) architecture with a minimal number of training epochs. The proposed model’s performance shows great results without a discriminator, a change from traditional GAN setups, thus emphasizing the need for simple and effective virtual try-on solutions. While the performance is characterized as good, this novel approach encourages further exploration in lightweight model design for practical and efficient try-on applications. Additionally, this paper showcases results from other GAN architectures providing a comprehensive overview of our research contributions in virtual try-on technology. Received: 10 January 2024 | Revised: 31 July 2024 | Accepted: 22 October 2024 Conflicts of Interest The authors declare that they have no conflicts of interest to this work. Data Availability Statement The data that support the findings of this study are openly available in the VITON-HD repository at https://www.dropbox.com/s/10bfat0kg4si1bu/zalando-hd-resized.zip?dl=0. The code for this research is available in the FashionFlow repository at https://github.com/noughtsamar/fashion-flow. Author Contribution Statement Samar Pratap: Conceptualization, Formal analysis, Data curation. Alston Richard Aranha: Methodology, Validation, Visualization. Divyanshu Kumar: Software, Investigation, Writing – original draft. Preethi P: Writing-review & editing, Supervision, Project administration.

Research Aims: This study aims to analyze various factors such as Perceived Usefulness, Perceived Ease of Use, Satisfaction, and Attitude in creating the intention to use Virtual Try-on technology in a sustainable manner for Somethinc make-up product users. Design/methodology/approach: This study uses quantitative approach with explanatory research. The research population was Mulawarman University students who experienced using Virtual Try-on. The sample was taken using multistage random sampling and amounted to 298 samples. Research Findings: The present study found that factors such as Confirmation, Perceived Ease of Use, Perceived Usefulness, Satisfaction, and Attitude affect Users' Continuance Intention in adopting Virtual Try-on technology to choose Somethinc make-up products. However, Attitude does not mediate the relationship between Perceived Usefulness and Continuance Intention, as well as between Perceived Ease of Use and Continuance Intention. The findings of this study suggest that positive attitudes are more easily formed due to ease of use, while usability is difficult to perceive as a trigger for continued use of technology. Theoretical Contribution/Originality: This study contributes to the existing literature by generating empirical evidence regarding the mediating role of Satisfaction and Attitude in shaping Continuance Intention. It emphasizes the need for Confirmation, Perceived Ease of Use, and Perceived Usefulness to enhance Satisfaction and Positive Attitude, which ultimately form Continuance Intention, especially in the use of Virtual Try-on. Keywords: Virtual Try-on; Confirmation; Perceived Ease of Use; Perceived Usefulness; Attitude; Satisfaction; Continuance Intention

The objectives of this study were to investigate whether images of female bodies shown in a website influence female consumers' level of body satisfaction and to examine how these variables affect online shoppers' concerns with garment fit and size and their intentions to use a virtual try-on model. We conducted an experiment using a 2 × 2 between-subject factorial design with 249 college students. The results showed the main effects of female bodies associated with body satisfaction on female consumers' concerns with garment fit and size. We also found a negative relationship between body satisfaction and concerns with garment fit and size as well as a positive relationship between concerns and intentions to use virtual try-on technology. However, we found no significant effect of female bodies on female consumers' body satisfaction.

Recently, the development of data communication networks and advances in computer processing capacity have significantly increased the amount of data to be managed. Although future innovations using data are expected, privacy violation of data has become a problem. For example, image disclosure in Social Networking Service may result in infringement of portrait rights. Previously, studies on anonymization have been conducted to disclose data for protecting personal information. However, conventional anonymization, even for high-dimensional data such as faced images, uses an averaging operation in anonymization and does not consider complex relationships between dimensions. Therefore, the loss of semantic meaning increases because it is caused by the anonymization process, assuming an Euclidean data space. We herein propose Multidimensional Inputs K-anonymizing Unit(MIKU), an anonymization algorithm focusing on face image anonymization as a typical example of high-dimensional datMIKU enables anonymization that retains the image quality. As the comparison method directly anonymizes images, the quality of images is deteriorated. MIKU considers the relation between dimensions using the latent space of StyleGAN. The effect of using latent space was confirmed by comparing the quantitative results of Frechet Inception Distance(FID) and the qualitative evaluation of the output image with the comparison method. The quantitative results by FID proved that MIKU achieved better performance when K-anonymity was large. When K-anonymity is small, we clarified that the image quality of the comparison method such as FID could not be measured correctly and was overestimated because of the effect of the inception model's linearity. In addition, the qualitative evaluation of anonymized images shows that MIKU generates a more natural image in the contour and hair expressions than the comparison method, and the anonymized images of MIKU contain no unnatural edge lines on a face such as those generated in the comparison method.

The recent success of Generative Adversarial Networks (GAN) is a result of their ability to generate high quality images given samples from a latent space. One of the applications of GANs is to generate images from a text description, where the text is first encoded and further used for the conditioning in the generative model. In addition to text, conditional generative models often use label information for conditioning. Hence, the structure of the meta-data and the ontology of the labels is important for such models. In this paper, we propose Ontology Generative Adversarial Networks (O-GANs) to handle the complexities of the data with label ontology. We evaluate our model on a dataset of fashion images with hierarchical label structure. Our results suggest that the incorporation of the ontology, leads to better image quality as measured by Frechet Inception Distance and Inception Score. Additionally, we show that the O-GAN better matches the generated images to their conditioning text, compared to models that do not incorporate the label ontology.

The virtual try-on technology based on 2D images aims to seamlessly transfer provided garments onto target person images. Prior methods mainly concentrated on warping garments and generating images, overlooking the influence of feature alignment on the try-on results. In this study, we initially analyze the distortions present by existing methods and elucidate the critical role of feature alignment in the extraction stage. Building on this, we propose a novel feature alignment-based model (FA-VTON). Specifically, FA-VTON aligns the upsampled higher-level features from both person and garment images to acquire precise boundary information, which serves as guidance for subsequent garment warping. Concurrently, the Efficient Channel Attention mechanism (ECA) is introduced to generate the final result in the try-on generation module. This mechanism enables adaptive adjustment of channel feature weights to extract important features and reduce artifact generation. Furthermore, to make the student network focus on salient regions of each channel, we utilize channel-wise distillation (CWD) to minimize the Kullback–Leibler (KL) divergence between the channel probability maps of the two networks. The experiments show that our model achieves better results in both qualitative and quantitative analyses compared to current methods on the popular virtual try-on datasets.

Virtual Try-on systems enable the try-on of a desired clothing on a target person image. These systems have led to vast research and have attracted commercial interest. However, the existing techniques are image-based systems limited to using an in-shop target clothing from a pre-defined dataset. To address this, we propose a video-based virtual try-on network DPV-VTON, that simulates the try-on using the target cloth extracted from the fashion videos on a target person image, while preserving the details and the characteristics. The core of the DPV-VTON pipeline is made up of (i) Best Frame Selection (BFS) module that extracts the best frame from the video (ii) Clothing Extraction module (CEM) extracts the target clothing from the selected best frame and generates a binary mask. (iii) A virtual try-on module synthesizes a final virtual try-on. Experiments on the existing benchmark datasets and a curated video dataset demonstrate that DPV-VTON generates photo-realistic and visually promising results. The proposed model obtains the lowest FID, LPIPS and the highest SSIM scores compared to the existing systems.

In this work, we propose a novel Cyclic Image Translation Generative Adversarial Network (CIT-GAN) for multi-domain style transfer. To facilitate this, we introduce a Styling Network that has the capability to learn style characteristics of each domain represented in the training dataset. The Styling Network helps the generator to drive the translation of images from a source domain to a reference domain and generate synthetic images with style characteristics of the reference domain. The learned style characteristics for each domain depend on both the style loss and domain classification loss. This induces variability in style characteristics within each domain. The proposed CIT-GAN is used in the context of iris presentation attack detection (PAD) to generate synthetic presentation attack (PA) samples for classes that are under-represented in the training set. Evaluation using current state-of-the-art iris PAD methods demonstrates the efficacy of using such synthetically generated PA samples for training PAD methods. Further, the quality of the synthetically generated samples is evaluated using Frechet Inception Distance (FID) score. Results show that the quality of synthetic images generated by the proposed method is superior to that of other competing methods, including StarGan.

BackgroundMalignant melanoma (MM) is the most aggressive skin cancer, requiring early diagnosis for better outcomes. While deep learning models have shown promise in dermatological image analysis, their performance is constrained by limited training data. Generative Adversarial Networks (GANs) offer a solution by generating synthetic images for data augmentation. However, assessing their clinical reliability remains difficult as automated metrics may not fully capture visual realism or clinical usability.The ObjectiveThis study presents a comprehensive framework for creating high-quality synthetic dermatoscopic pictures of MM lesions, as well as a holistic validation methodology that uses quantitative metrics and qualitative dermatologist assessment to provide a full clinical evaluation of the generated images. Three GAN architectures (DCGAN, StyleGAN2, and StyleGAN3-t) are explored. Lesions on the face, palms, and soles are excluded due to their unique dermoscopic patterns.Materials and MethodsA dataset of 1,774 dermatoscopic body MM images was used to train the models, assessed with Fréchet Inception Distance (FID), Kernel Inception Distance (KID), precision, and recall. Afterwards, a panel of 17 dermatologists with different levels of expertise assessed image quality using a 7-point Likert scale, with accuracy, sensitivity, specificity, and inter-rater agreement analysed.ResultsStyleGAN2 achieved the best image fidelity (FID: 18.89, KID: 0.0025), while StyleGAN3-t demonstrated stable but slower convergence. Both StyleGAN models outperformed DCGAN in diversity and fidelity. The validation study showed that StyleGAN2-generated images were often indistinguishable from real ones, reflected in low specificity and accuracy values among evaluators.ConclusionsThe study highlights the effectiveness of GANs in generating high-quality synthetic images, proposing a validation framework that integrates expert assessments with state-of-the-art quantitative metrics. This approach advances standardisation in GAN evaluation, ensuring synthetic images are clinically relevant for dermatological AI applications.

Megavoltage computed tomography (MVCT) is an essential imaging modality for verifying patient positioning in helical tomotherapy. However, its clinical application in daily anatomical monitoring and adaptive radiotherapy is hindered by inherent image artifacts and poor soft-tissue contrast. This issue is particularly pronounced in pelvic radiotherapy, where the intra- and interfraction anatomical variations necessitate high-quality image guidance to ensure precise dose delivery. We developed a deep-learning-based framework for MVCT enhancement to improve anatomicalvisualizationand facilitate accurate adaptive treatment planning for cervical cancer. This study analyzed a retrospective cohort of 170 patients with cervical cancer who underwent helical tomotherapy. The proposed deep-learning-based algorithm employed a generative adversarial network (GAN) that integrated deformable convolution and a self-attention mechanism (SADC-EGAN) to improve MVCT image quality. Comparative analyses were conducted against representative baseline methods, including U-Net, Attention U-Net, U-Net++, Swin-UNet, CycleGAN, and Pix2Pix. Model performance was assessed using quantitative metrics, including mean absolute error (MAE), peak signal-to-noise ratio (PSNR), structural similarity index measure (SSIM), and Fréchet inception distance (FID). The synthetic computed tomography (sCT) images generated by the proposed SADC-EGAN method demonstrated superior Hounsfield unit (HU) accuracy and structural similarity comparedto the original MVCT. Specifically, the MAE between the sCT and kilovoltage computed tomography (kVCT) was reduced to 36.32±6.69 HU, compared with 56.72±9.09 HU for MVCT. In terms of image quality, the sCT images exhibited notable enhancements over MVCT images, with higher PSNR (32.54±2.31 vs. 29.40±1.56dB), improved SSIM (0.93±0.01vs. 0.89±0.02), and substantially lower FID (66.68±22.31vs. 153.52±28.77). The proposed SADC-EGAN framework, integrating deformable convolutions and self-attention, effectively generated high-quality kVCT-like images from MVCT, improving both HU accuracy and image quality. This approachhas clinical potential to enable online adaptive helical tomotherapy for cervical cancer.

Existing image-based virtual try-on methods often require personal photos, causing privacy violations. We thus propose a novel Multi-Scene Virtual Try-on Network (MSVTON), which generates try-on images while only giving clothes images and attribute texts consisting of view angle, skin color, etc. With our model, various scene try-on images can be created by manipulating the attributes. In order to improve semantic consistency and visual realism of the generated images, MSVTON decomposes the virtual try-on process into two stages. MS-VTON consists of Scene Learning Network (SLN) and Content Learning Network (CLN). SLN learns the semantics of try-on scenes and generates coarse try-on images in the first stage, and CLN refines minutiae such as clothes textures, etc. in the second stage. Experiments show that the FID score of MSVTON reaches 9. S, which outperforms the existing image-toimage translation methods.

Free-VTON: Cost-free acceleration and quality enhancement for diffusion-based virtual try-on.

One of the major challenges in medical imaging is the limited availability of high-quality datasets. To address this, Generative Artificial Intelligence (Generative AI) offers a promising solution by generating synthetic medical images to augment existing datasets. This study explores the application of Deep Convolutional Generative Adversarial Networks (DCGAN) for data augmentation in CIN imaging. Two training scenarios were implemented: DCGAN with manual data augmentation and another without manual augmentation. The image quality was evaluated using the Fréchet Inception Distance (FID). The results indicate that incorporating data augmentation improves the stability of training and enhances the quality of generated images FID scores of 2.21. In contrast, training DCGAN without manual augmentation resulted in a higher FID score of 2.52, indicating lower image quality. These findings highlight the effectiveness of DCGAN in medical image augmentation and its potential to enhance deep learning-based diagnostic models for cervical cancer detection or classification.

Cone-beam CT (CBCT) is widely used in dental clinics but exhibits limitations in assessing soft tissue pathology because of its lack of contrast resolution and low Hounsfield Units (HU) quantification accuracy. We aimed to increase the image quality and HU accuracy of CBCTs while preserving anatomical structures. We generated CT-like images from CBCT images using a patchwise contrastive learning-based GAN model. Our model was trained on unpaired CT and CBCT datasets with the novel combination of losses and the feature extractor pretrained on our training dataset. We evaluated the quality of the images generated by our model in terms of Fréchet inception distance (FID), peak signal-to-noise ratio (PSNR), mean absolute error (MAE), and root mean square error (RMSE). Additionally, the structure preservation performance was assessed by the structure score. As a result, the generated CT-like images by our model were significantly superior to those generated by various baseline models in terms of FID, PSNR, MAE, RMSE, and structure score. Therefore, we demonstrated that our model provided the complementary benefits of preserving the anatomical structures of the input CBCT images and improving the image quality to be similar to those of CT images.