A Deep Learning Approach to Interface Color Quality Assessment in HCI

Jiacheng Hu; Junliang Du; Ran Hao; Runsheng Zhang; Shixiao Wang

REVIEW 4 major objections 4 minor 22 references

A Deep Learning Approach to Interface Color Quality Assessment in HCI

T0 review · 4 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read A convolutional network can turn interface screenshots into color-quality scores that closely track human ratings, with Pearson correlations up to 0.96 on diverse website designs.

desk verdict Table 1 is internally inconsistent, so the paper's central claim of 0.96 agreement with user ratings cannot be correct as reported; the method is also too under-specified to evaluate or reproduce. read the letter →

arxiv 2502.09914 v1 pith:ZGXB4P22 submitted 2025-02-14 cs.HC

Shixiao Wang , Runsheng Zhang , Junliang Du , Ran Hao , Jiacheng Hu This is my paper · ORCID

classification cs.HC

keywords interfacecolorqualitydeepconvolutionalneuralnetworkassessmenthuman-computerinteractionuserratingcorrelationhuelightnesspuritydesignevaluationquantitative

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper tries to establish that a convolutional neural network can turn an interface screenshot into quantitative color-quality scores—contrast, clarity, color coordination, and visual appeal—that closely track what human raters say. The model combines color-distribution statistics with CNN-extracted high-level features and learns a weighted quality score from user ratings. On screenshots drawn from e-commerce, social media, education, news, and cloud-service sites, the reported Pearson correlations with user ratings reach 0.96 for color coordination and stay above 0.93 for the other three metrics. If the claim holds, interface evaluation could move from designer intuition to a reproducible, data-driven measurement, giving design teams a fast way to test and optimize color choices.

What carries the argument

The load-bearing mechanism is a deep convolutional neural network—a deep image model that learns hierarchical visual features—read out by a weighted quality formula. The input screenshot is resized to a fixed $H_0 \times W_0 \times 3$ tensor and passed through convolutional layers to produce feature maps $F_l$; the deepest map is globally average-pooled while shallow maps contribute per-channel means and standard deviations, and these are concatenated into a vector $z$. From $z$ the model produces quality components for hue, lightness, and purity: hue quality uses the entropy $-\sum_i p_i \log p_i$ of the HSV hue histogram, lightness and purity use the ratio of standard deviation to mean, and the components are combined linearly into the final score, trained by mean squared error against user ratings. This design lets the network supply learned perceptual features while the explicit color-statistics terms keep the output interpretable as a color-quality score.

What would settle it

Take the reported dataset, partition it with an explicit train/test split at the website level, retrain the CNN, and compute the four Pearson correlations on the held-out screenshots; if they fall substantially below the reported 0.93–0.96 range, the claimed agreement reflects fit to training data rather than predictive agreement with users.

Watch

Extended reading notes

Core claim

The central claim is that interface color quality is a learnable function of hue, lightness, and purity structure, and that a deep CNN approximates that function closely enough to serve as a data-driven alternative to subjective evaluation. The model's output is a weighted sum $Q = \alpha Q_{\mathrm{hue}} + \beta Q_{\mathrm{lightness}} + \gamma Q_{\mathrm{purity}}$, with $\alpha + \beta + \gamma = 1$; each component is derived from CNN feature maps, with hue quality measured by the entropy of the hue distribution and lightness/purity qualities by their coefficient of variation. The network is trained to minimize mean squared error against human-assigned scores. The paper reports that CNN scores and user ratings align with Pearson correlations of 0.94 (contrast), 0.93 (clarity), 0.96 (color coordination), and 0.95 (visual appeal), with small MSE and MAE values.

Load-bearing premise

The load-bearing assumption is that the Table 1 correlations were computed on interface images the network had not seen during training; the paper does not state that any held-out split exists, so the strength of the claimed agreement with user ratings depends on that unstated condition.

Editorial extensions

If this is right

Interface design teams could obtain per-metric color scores for contrast, clarity, coordination, and appeal without recruiting a large user panel for every design iteration.
The model could serve as an automated objective function, letting designers or generative tools tune color schemes to maximize predicted quality scores.
Color quality becomes auditable and comparable across websites and design styles, because the same learned metric is applied to every screenshot.
Failures in hue or brightness balance can be diagnosed from the component sub-scores rather than from a single holistic rating.
The same CNN-based evaluation could be retrained on other visual interface types, such as smart-home, medical, or virtual-reality interfaces, as the paper identifies for future work.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Editorial inference: if the reported correlations are reproduced on a clearly held-out set, the model's strongest practical role is as a screening tool that flags likely color-quality problems before human testing, not as a complete substitute for final human judgment.
Editorial inference: the paper's observation that strong contrast peaks near 180° on the color wheel suggests a rotational symmetry prior could be built into the network's color features, potentially improving predictions on new sites.
Editorial inference: a natural test is to compare this CNN against a non-deep baseline using only hue entropy and lightness/purity coefficient-of-variation features; that comparison would reveal how much the learned features add over the explicit color statistics.
Editorial inference: because the dataset spans culturally distinct websites, grouping user ratings by site region could test whether the model's agreement with users holds equally across design cultures or whether it absorbs a specific aesthetic bias.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The headline is this: the main quantitative result is not just weakly supported, it is internally contradicted by the paper's own table. The stress-test note is correct and decisive. For paired evaluations, the mean absolute error must be at least as large as the absolute difference of the means. Table 1 violates this for every row: contrast |0.74-0.76|=0.02 but MAE=0.019; clarity 0.03>0.018; color coordination 0.03>0.014; visual appeal 0.04>0.016. That means the reported averages, the MAE, or the column definitions are wrong. This is a fatal problem for the paper's central claim, not a quibble. The reader's take is otherwise accurate. The paper is a standard CNN regression on website screenshots, with no architecture, no dataset size, no train/validation split, no baselines, and no comparison to the extensive image aesthetic assessment literature. The equations in Section III are garbled, and alpha, beta, gamma are never reported, so the claimed weighted-sum quality model is not actually specified. The high correlations are presented without any indication that they come from held-out data; if they are computed on the training set, they are circular by construction. What can I credit? The problem is real and worth working on: quantifying interface color quality with a combination of hand-designed color statistics and learned features is a sensible direction. The idea of collecting screenshots across e-commerce, social media, and education sites and comparing CNN scores against a 500-user evaluation is also not crazy. But the execution does not back these ambitions. I see no new technical contribution here, and the evidence presented cannot support the abstract's claims. The citation pattern is mostly to a cluster of self-citations and unrelated deep learning papers; that is not by itself a flaw, but it does not help the reader understand where this fits. No code or data is shipped, so there is nothing independently checkable. Who is this for? A reader looking for a survey of HCI color evaluation methods might find the problem statement useful, but no one should rely on the numbers. This does not deserve a serious referee at this stage. My recommendation: reject and do not invite resubmission unless the authors completely rewrite the experimental section, fix Table 1, release the dataset and code, and compare against standard baselines. As it stands, the paper is not coherent enough to review meaningfully.

Referee Report

4 major / 4 minor

Summary. The paper proposes a deep convolutional network (CNN) for quantitative assessment of interface color quality in HCI. The method combines handcrafted hue, lightness, and purity statistics with CNN-extracted features, trains on manually annotated quality scores via a squared-error loss, and then reports high agreement with 500-user ratings on contrast, clarity, color coordination, and visual appeal, with Pearson correlations up to 0.96 and low MSE/MAE values. The central contribution is the claimed consistency between CNN predictions and human subjective ratings, supported mainly by Table 1 in Section IV-B.

Significance. If the claimed correlations were obtained on held-out data, an automated interface color quality predictor matching human ratings would be practically useful for HCI design and evaluation. The topic is relevant, but the manuscript provides no code, data, machine-checked proofs, or reproducible protocol. Its significance is entirely dependent on the validity of the experimental evidence, which the paper does not currently establish.

major comments (4)

[IV-B, Table 1] Table 1 is internally inconsistent. If the columns are what they appear to be—the mean user rating and mean CNN rating on the same interfaces—then MAE = mean_i |u_i - c_i| is bounded below by |mean_i u_i - mean_i c_i|. Each row violates this inequality: contrast has |0.74 - 0.76| = 0.02 > 0.019; clarity has |0.72 - 0.75| = 0.03 > 0.018; color coordination has |0.70 - 0.73| = 0.03 > 0.014; visual appeal has |0.73 - 0.77| = 0.04 > 0.016. Therefore the reported average ratings and MAE values cannot have been computed from the same paired data. Since Table 1 is the central experimental evidence, the paper's main claim is unsupported as presented.
[III and IV-B] The training and evaluation protocol is circular as described. Section III defines the loss L = (1/N) * sum_i (y_i - Q(i; theta))^2 with y_i a manually annotated quality score, but Section IV-B reports Pearson, MSE, and MAE between CNN scores and those same user ratings without stating any train/validation/test split. Without a held-out set, high correlation is expected from fitting the training data. The paper must specify how many interfaces were rated, how the split was made, and report metrics on a held-out test set only.
[III] The method section is too underspecified to support the claims. No CNN architecture is given (number of layers, kernel sizes, pooling strategy, activation functions), no input size W0 x H0, no training hyperparameters (optimizer, epochs, batch size), and no dataset size per category. In addition, the weight coefficients alpha, beta, gamma in Q = alpha*Q_hue + beta*Q_lightness + gamma*Q_purity are never reported, and the number of hue intervals n is not specified. These details are required to reproduce the experiments and to interpret the reported accuracy.
[IV-B, Figure 4] The reported improvements in balance, continuity, integrity, and unity after 'CNN training' are not evidence for the model's validity. The CNN was trained to approximate user ratings, so applying it and then comparing metrics of its own outputs to the original images is likely circular. There is no description of how those metrics were computed or on which samples. This claim should be removed or replaced with an independent evaluation on data not used for training.

minor comments (4)

[I and Abstract] The phrase 'avoid the influence of subjective factors' is overstated because the model is trained on subjective user ratings; please soften or qualify the claim.
[III] The equations in Section III contain garbled symbols (e.g., the tensor product, the feature map definition, and the loss function), and the figure references are not clearly integrated. The text needs careful proofreading.
[References] Several references appear only tangentially related to the claims (e.g., [4], [6]-[8]); revising the reference list to focus on prior work on aesthetic quality assessment and color perception would improve the paper.
[V] The conclusion claims generalization ability, but no cross-validation or external dataset evaluation is reported; the conclusion should be limited to what the experiments actually support.

Circularity Check

2 steps flagged · score 7.0 of 10

The claimed CNN–user agreement is the training objective: the model is fit to user quality scores and the same scores are then reported as prediction accuracy, with no stated held-out split.

fitted input called prediction [Section III (Method), supervised-learning paragraph; Section IV.B, Table 1]
"Finally, we use supervised learning to train the evaluation model. Given an interface image dataset {l_i, y_i}_{i=1}^N, where y_i is the manually annotated quality score of the i-th interface, the loss function is defined as: L = 1/N Σ_i (Q(l_i; θ) − y_i)^2 ... By optimizing L, we learn the optimal model parameters. Table 1 Consistency analysis between CNN evaluation and user reviews."

The network parameters (including the α, β, γ weights) are optimized by minimizing squared error against y_i, the quality scores. Table 1 then reports Pearson, MSE, and MAE between the CNN's scores and the user ratings of 'the same interfaces.' No train/validation/test split is described anywhere in the paper, so as written the reported MSE is the same squared-error objective that was minimized during training. High correlation and low error on the fitting labels are forced by the loss and do not demonstrate that the model predicts anything beyond its own training target.
self definitional [Section III (Method), weighted-sum quality model]
"Based on the aesthetic evaluation research in psychology, the evaluation of interface quality can be expressed as a weighted sum model: Q = α·Q_hue + β·Q_lightness + γ·Q_purity, where Q_hue, Q_lightness and Q_purity represent the quality scores of hue, lightness and purity respectively, and α, β and γ are the corresponding weight coefficients, satisfying α+β+γ=1."

The paper never specifies the values of α, β, γ or an independent rule for setting them; the only optimization procedure stated for the whole model is the supervised loss L = 1/N Σ (Q(l_i; θ) − y_i)^2 against the quality scores. The 'quality score' Q is therefore defined as the function that best reproduces the user ratings, and the later 'consistency analysis' between CNN and user ratings is comparing the definition with its own fitting target.

full rationale

The central quantitative claim is that the CNN evaluation is highly consistent with user ratings (Pearson up to 0.96). But the method section defines the model's training loss as squared error against manually annotated quality scores, and the experiment section reports agreement with user ratings on 'the same interfaces' without ever stating that a held-out set was used. In the absence of any split, the reported MSE/MAE/correlation are the training objective renamed as an evaluation result; this is fitted input called prediction. The weighted-sum definition of Q with unreported α, β, γ reinforces the circularity because those coefficients are part of the parameters fitted to the same labels. Separately, Table 1 is internally inconsistent as printed: for paired means, MAE ≥ |mean_user − mean_CNN|, yet every row violates this (e.g., Contrast |0.74−0.76| = 0.02 > MAE 0.019; Visual appeal |0.73−0.77| = 0.04 > MAE 0.016), so the numbers cannot be a faithful record of the claimed comparison. These issues affect the paper's core contribution; no external benchmark or independent test set is offered to break the circularity.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The central claim depends on several unstated choices: the linear quality weights, the CNN architecture and its fitted parameters, the number of hue bins, and the evaluation split. None of these are reported, so the reader cannot separate the model's genuine predictive power from fitted quantities.

free parameters (4)

alpha, beta, gamma = Not reported.
Weights in Q = alpha * Q_hue + beta * Q_lightness + gamma * Q_purity are constrained to sum to 1 but their fitted values are never stated. The final score depends on them.
CNN weights and bias parameters = Not reported.
All convolutional and fully connected weights are optimized against user ratings via the mean squared error loss. No architecture or parameter count is given.
Hue entropy interval count n = Not reported.
The entropy calculation divides hue values into n intervals. The choice of n affects Q_hue and is not justified or specified.
Contrast angle model coefficients = Not reported.
Figure 3 fits a model of contrast intensity versus color-wheel angle. The fitted coefficients are not given, so the model cannot be checked or reused.

assumptions (4)

domain assumption Interface color quality is decomposable into hue, lightness, and purity components with a weighted linear sum.
Invoked in Section III in the definition Q = alpha * Q_hue + beta * Q_lightness + gamma * Q_purity. No empirical or theoretical justification is given for linear separability.
domain assumption Manual user ratings are valid ground truth for color quality.
The loss function in Section III treats y_i as ground truth. The paper does not report inter-annotator agreement or control for cultural and individual variation.
domain assumption CNNs can extract perceptual color quality features from raw pixels.
The whole method assumes deep features correlate with aesthetic judgment. This is plausible from prior work but is not established here beyond the reported correlations.
domain assumption Scraped public website homepages form a representative dataset for interface color quality.
Section IV-A lists website categories but gives no counts, sampling rules, or annotation protocol.

how reviews work

0 comments

Cite this review

Pith. "Pith review of A Deep Learning Approach to Interface Color Quality Assessment in HCI." pith.science (2026). https://pith.science/paper/ZGXB4P22

@misc{pith2026250209914,
  author       = {Pith},
  title        = {Pith review of: A Deep Learning Approach to Interface Color Quality Assessment in HCI},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZGXB4P22}},
  note         = {Machine review of arXiv:2502.09914}
}

read the original abstract

In this paper, a quantitative evaluation model for the color quality of human-computer interaction interfaces is proposed by combining deep convolutional neural networks (CNN). By extracting multidimensional features of interface images, including hue, brightness, purity, etc., CNN is used for efficient feature modeling and quantitative analysis, and the relationship between interface design and user perception is studied. The experiment is based on multiple international mainstream website interface datasets, covering e-commerce platforms, social media, education platforms, etc., and verifies the evaluation effect of the model on indicators such as contrast, clarity, color coordination, and visual appeal. The results show that the CNN evaluation is highly consistent with the user rating, with a correlation coefficient of up to 0.96, and it also shows high accuracy in mean square error and absolute error. Compared with traditional experience-based evaluation methods, the proposed model can efficiently and scientifically capture the visual characteristics of the interface and avoid the influence of subjective factors. Future research can explore the introduction of multimodal data (such as text and interactive behavior) into the model to further enhance the evaluation ability of dynamic interfaces and expand it to fields such as smart homes, medical systems, and virtual reality. This paper provides new methods and new ideas for the scientific evaluation and optimization of interface design.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

22 extracted references · 20 canonical work pages

[1]

Quantitative and Qualitative Analysis of Interaction Design and Usability: A Bibliometric Approach,

Q. Shan, Y. M. Yusoff and A. A. Mutalib, “Quantitative and Qualitative Analysis of Interaction Design and Usability: A Bibliometric Approach,” Kurdish Studies, vol. 12, no. 2, pp. 2752-2765, 2024

work page 2024
[2]

Research on Evaluation System of Human-Computer Interaction System in University Library Under the Background of Artificial Intelligence,

H. Q. Dong, G. H. Cao and L. Huang, “Research on Evaluation System of Human-Computer Interaction System in University Library Under the Background of Artificial Intelligence,” Proceedings of the 2024 6th Asia Pacific Information Technology Conference, pp. 69-77, 2024

work page 2024
[3]

Evaluation of a Smart Audio System Based on the ViP Principle and the Analytic Hierarchy Process Human– Computer Interaction Design,

J. Huang, W. Li and T. Sadad, “Evaluation of a Smart Audio System Based on the ViP Principle and the Analytic Hierarchy Process Human– Computer Interaction Design,” Applied Sciences, vol. 14, no. 7, p. 2678, 2024

work page 2024
[4]

Adversarial Attack Against Images Classification based on Generative Adversarial Networks,

Y. Yang, "Adversarial Attack Against Images Classification based on Generative Adversarial Networks," arXiv preprint arXiv:2412.16662, 2024

arXiv 2024
[5]

Scaling-up medical vision-and- language representation learning with federated learning,

S. Lu, Z. Liu, T. Liu, and W. Zhou, "Scaling-up medical vision-and- language representation learning with federated learning," Engineering Applications of Artificial Intelligence, vol. 126, p. 107037, 2023

work page 2023
[6]

Comprehensive Evaluation of Multimodal AI Models in Medical Imaging Diagnosis: From Data Augmentation to Preference-Based Comparison

C. Ruan, C. Huang, and Y. Yang, "Comprehensive Evaluation of Multimodal AI Models in Medical Imaging Diagnosis: From Data Augmentation to Preference-Based Comparison," arXiv preprint, arXiv:2412.05536, 2024

work page Pith review arXiv 2024
[7]

Improving the RAG- based Personalized Discharge Care System by Introducing the Memory Mechanism

Y. Yang, C. Xu, J. Guo, T. Feng, and C. Ruan, "Improving the RAG- based Personalized Discharge Care System by Introducing the Memory Mechanism", Preprints, doi: 10.20944/preprints202410.1696.v1, 2024

work page arXiv 2024
[8]

A Self-training Framework for Automated Medical Report Generation,

S. Wang, Z. Liu and B. Peng, "A Self-training Framework for Automated Medical Report Generation," Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pp. 16443-16449, December 2023

work page 2023

Show all 22 references

[9]

Gesture-Driven Innovation: Exploring the Intersection of Human-Computer Interaction and Virtual Fashion Try-On Systems,

M. Prakash, N. Arunkumar and G. Gokul, “Gesture-Driven Innovation: Exploring the Intersection of Human-Computer Interaction and Virtual Fashion Try-On Systems,” 2024 2nd International Conference on Networking and Communications (ICNWC), pp. 1-10, 2024

2024
[10]

Human-Computer Interaction Approach with Empathic Conversational Agent and Computer Vision,

R. Pereira, C. Mendes, N. Costa et al., “Human-Computer Interaction Approach with Empathic Conversational Agent and Computer Vision,” International Work-Conference on the Interplay Between Natural and Artificial Computation, Cham: Springer Nature Switzerland, pp. 431- 440, 2024

2024
[11]

Systematic Analysis of User Perception for Interface Design Enhancement,

S. Duan, “Systematic Analysis of User Perception for Interface Design Enhancement,” J. Comput. Sci. Softw. Appl., vol. 5, no. 2, 2024

2024
[12]

Adaptive User Interface Generation Through Reinforcement Learning: A Data-Driven Approach to Personalization and Optimization,

Q. Sun, Y. Xue, and Z. Song, “Adaptive User Interface Generation Through Reinforcement Learning: A Data-Driven Approach to Personalization and Optimization,” arXiv preprint arXiv:2412.16837, 2024

2024 arXiv
[13]

Deep Learning-Based Gesture Key Point Detection for Human-Computer Interaction Applications,

S. Duan, “Deep Learning-Based Gesture Key Point Detection for Human-Computer Interaction Applications,” Trans. Comput. Sci. Methods, vol. 5, no. 1, 2025

2025
[14]

Computer Vision- Driven Gesture Recognition: Toward Natural and Intuitive Human- Computer,

F. Shao, T. Zhang, S. Gao, Q. Sun, and L. Yang, “Computer Vision- Driven Gesture Recognition: Toward Natural and Intuitive Human- Computer,” arXiv preprint arXiv:2412.18321, 2024

2024 arXiv
[15]

Optimized Convolutional Neural Network for Intelligent Financial Statement Anomaly Detection,

X. Du, “Optimized Convolutional Neural Network for Intelligent Financial Statement Anomaly Detection,” J. Comput. Technol. Softw., vol. 3, no. 9, 2024

2024
[16]

Mining Multimodal Data with Sparse Decomposition and Adaptive Weighting,

X. Wang, “Mining Multimodal Data with Sparse Decomposition and Adaptive Weighting,” Trans. Comput. Sci. Methods, vol. 5, no. 1, 2025

2025
[17]

Multi-Level Attention and Contrastive Learning for Enhanced Text Classification with an Optimized Transformer,

J. Gao, G. Liu, B. Zhu, S. Zhou, H. Zheng, and X. Liao, “Multi-Level Attention and Contrastive Learning for Enhanced Text Classification with an Optimized Transformer,” arXiv preprint arXiv:2501.13467, 2025

2025 arXiv
[18]

Multivariate Time Series Forecasting and Classification via GNN and Transformer Models,

J. Wang, “Multivariate Time Series Forecasting and Classification via GNN and Transformer Models,” J. Comput. Technol. Softw., vol. 3, no. 9, 2024

2024
[19]

Dynamic Distributed Scheduling for Data Stream Computing: Balancing Task Delay and Load Efficiency,

X. Sun, “Dynamic Distributed Scheduling for Data Stream Computing: Balancing Task Delay and Load Efficiency,” J. Comput. Technol. Softw., vol. 4, no. 1, 2025

2025
[20]

Improved Transformer for Cross-Domain Knowledge Extraction with Feature Alignment,

P. Li, “Improved Transformer for Cross-Domain Knowledge Extraction with Feature Alignment,” J. Comput. Sci. Softw. Appl., vol. 5, no. 2, 2024

2024
[21]

Pietschmann, Human-Computer Interaction in Extended Reality: Exploring the Impact of Visual Guidance on User Performance and Human Factors, 2024

L. Pietschmann, Human-Computer Interaction in Extended Reality: Exploring the Impact of Visual Guidance on User Performance and Human Factors, 2024

2024
[22]

Research Trends of Human-Machine Interaction Studies in Mechanical Equipment Design: A Bibliometric Review,

Y. Wang, Y. Gan, Y. Li et al., “Research Trends of Human-Machine Interaction Studies in Mechanical Equipment Design: A Bibliometric Review,” International Conference on Human-Computer Interaction, Cham: Springer Nature Switzerland, pp. 436-443, 2024

2024

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

Quantitative and Qualitative Analysis of Interaction Design and Usability: A Bibliometric Approach,

Q. Shan, Y. M. Yusoff and A. A. Mutalib, “Quantitative and Qualitative Analysis of Interaction Design and Usability: A Bibliometric Approach,” Kurdish Studies, vol. 12, no. 2, pp. 2752-2765, 2024

work page 2024

[2] [2]

Research on Evaluation System of Human-Computer Interaction System in University Library Under the Background of Artificial Intelligence,

H. Q. Dong, G. H. Cao and L. Huang, “Research on Evaluation System of Human-Computer Interaction System in University Library Under the Background of Artificial Intelligence,” Proceedings of the 2024 6th Asia Pacific Information Technology Conference, pp. 69-77, 2024

work page 2024

[3] [3]

Evaluation of a Smart Audio System Based on the ViP Principle and the Analytic Hierarchy Process Human– Computer Interaction Design,

J. Huang, W. Li and T. Sadad, “Evaluation of a Smart Audio System Based on the ViP Principle and the Analytic Hierarchy Process Human– Computer Interaction Design,” Applied Sciences, vol. 14, no. 7, p. 2678, 2024

work page 2024

[4] [4]

Adversarial Attack Against Images Classification based on Generative Adversarial Networks,

Y. Yang, "Adversarial Attack Against Images Classification based on Generative Adversarial Networks," arXiv preprint arXiv:2412.16662, 2024

arXiv 2024

[5] [5]

Scaling-up medical vision-and- language representation learning with federated learning,

S. Lu, Z. Liu, T. Liu, and W. Zhou, "Scaling-up medical vision-and- language representation learning with federated learning," Engineering Applications of Artificial Intelligence, vol. 126, p. 107037, 2023

work page 2023

[6] [6]

Comprehensive Evaluation of Multimodal AI Models in Medical Imaging Diagnosis: From Data Augmentation to Preference-Based Comparison

C. Ruan, C. Huang, and Y. Yang, "Comprehensive Evaluation of Multimodal AI Models in Medical Imaging Diagnosis: From Data Augmentation to Preference-Based Comparison," arXiv preprint, arXiv:2412.05536, 2024

work page Pith review arXiv 2024

[7] [7]

Improving the RAG- based Personalized Discharge Care System by Introducing the Memory Mechanism

Y. Yang, C. Xu, J. Guo, T. Feng, and C. Ruan, "Improving the RAG- based Personalized Discharge Care System by Introducing the Memory Mechanism", Preprints, doi: 10.20944/preprints202410.1696.v1, 2024

work page arXiv 2024

[8] [8]

A Self-training Framework for Automated Medical Report Generation,

S. Wang, Z. Liu and B. Peng, "A Self-training Framework for Automated Medical Report Generation," Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pp. 16443-16449, December 2023

work page 2023