eISSN : 3022-7011
ISSUER : KIPS
 
After the Korea Information Processing Society (KIPS) Transactions journal was founded in 1994, it was reorganized into the KIPS Transactions: Computer and Communication Systems(2287-5891/2734-049X ) and the KIPS Transactions: Software and Data Engi neering(2287-5905/2734-0503) in 2012. Through the KIPS official meeting on January 8th, 2024, the new KIPS Transaction journal was founded by integrating two KIPS Journals, KIPS Transactions: Computer and Communication Systems and KIPS Transactions: Software and Data Engineering. The new journal aims to realize social value and contribute to the development of South Korea’s science and technology with support from the lottery fund of the Ministry of Strategy and Finance and the science/technology promotion fund of the Ministry of Science and ICT. It is indexed in the Korea Science Academic Database, Korea Citation Index (KCI), and EBSCO.

HighlightsMore

Diffusion-based Audio-to-Visual Generation for High-Quality Bird Images

Adel Toleubekova  Joo Yong Shim  XinYu Piao  Jong-Kook Kim
Accurately identifying bird species from their vocalizations and generating corresponding bird images is still a challenging task due to limited training data and environmental noise in audio data. To address this limitation, this paper introduces a...

Bambda: A Framework for Preventing Function Invocation Condition-Based Attacks in Serverless Environments

Shin Chang Hee  Lee Seung Soo
Serverless computing is rapidly emerging as a new paradigm in cloud computing, offering automatic scalability, cost efficiency, and ease of operation. However, its two core characteristics—IAM-based privilege management and event-driven execution—ca...

PEFT Methods for Domain Adaptation

Lee You Jin  Yoon Kyung Koo  Chung Woo Dam
This study analyzed that the biggest obstacle in deploying Large Language Models (LLMs) in industrial settings is incorporating domain specificity into the models. To mitigate this issue, the study compared model performance when domain knowledge wa...

Label Differential Privacy Study for Privacy Protection in Multimodal Contrastive Learning Model

Youngseo Kim  Minseo Yu  Younghan Lee  Ho Bae
Recent advancements in multimodal deep learning have garnered significant attention from both academia and industry due to their exceptional accuracy and ability to learn rich knowledge representations. In particular, contrastive learning based appr...

Latest Publication   (Vol. 14, No. 8, Aug.  2025)

Research on Exploiting Browser Use-After-Free Bugs Triggered by User Interaction
Youngjoo Lee  Byoungyoung Lee
This study demonstrates that browser UI (User Interface) bugs triggered by user interactions can pose a security threat and introduces a method to exploit them to attack the main browser process. Web browsers are essential software used by hundreds of millions, and vulnerabilities in them can lead to large-scale security incidents if exploited. While UI bugs have traditionally been regarded as minor usability issues with lower patch priority, this study designs an attack scenario that leverages memory corruption vulnerabilities, such as Use-After-Free (UAF), caused by UI bugs to manipulate browser execution flow and conducts experiments using actual vulnerabilities. Through this, the study aims to raise awareness of the security risks posed by UI bugs in browsers and highlight the necessity of addressing such vulnerabilities
The Transactions of the Korea Information Processing Society, Vol. 14, No. 8, pp. 569-575, Aug. 2025
https://doi.org/10.3745/TKIPS.2025.14.8.569
Web Browser User Interface Vulnerability Attack
A Comparative Study of Feature Importance Algorithms and Feature Selection for Static Feature-Based Ransomware Detection
Jeon Hye Min  Choi Doo Seop  Im Eul Gyu
In this paper, we extract 54 static features from ransomware PE files—including header metadata, section sizes, and virtual memory sizes—and evaluate their importance using four algorithms: Gain Ratio, Information Gain, Gini Importance, and Mutual Information. For each algorithm, we select the top-K features to form a reduced feature set, which is then used to train and validate four classification models: Random Forest, Decision Tree, Support Vector Machine, and Multi-Layer Perceptron. Experimental results show that the Random Forest model, using 41 features selected by a Gain Ratio threshold of K = 0.01, achieves the highest accuracy of 99.33%. The Decision Tree, SVM, and MLP models also demonstrate strong performance with accuracies of 98.67%, 96.67%, and 98.75%, respectively. These findings confirm that careful feature selection can substantially reduce computational costs while maintaining high detection accuracy
The Transactions of the Korea Information Processing Society, Vol. 14, No. 8, pp. 576-587, Aug. 2025
https://doi.org/10.3745/TKIPS.2025.14.8.576
Ransomware feature importance Static Feature Machine Learning
Design and Implementation of A Laundry Folding Automation System Based on Deep Learning
Kim Eun Sun  Son Jin-Gon
This study designed an automated system to improve household labor efficiency by automating clothing recognition and folding tasks. A CNN model was trained on the Fashion MNIST dataset to classify tops, and an OV7670 camera sensor was used to recognize clothing. Based on the classification results, the folding points were predicted, and an algorithm was implemented to automate the folding process. The study found that the CNN model effectively classified the clothing, and the folding algorithm successfully performed tasks according to preset rules. These findings suggest that integrating computer vision and automation can significantly reduce household labor, providing a foundation for future smart home applications.
The Transactions of the Korea Information Processing Society, Vol. 14, No. 8, pp. 588-594, Aug. 2025
https://doi.org/10.3745/TKIPS.2025.14.8.588
clothing recognition CNN Fashion MNIST Automation Smart Home
Intrusion Detection-Based Access Control Algorithm for PostgreSQL Based on Front-End Back-End Protocols
Seong-Hoon Jwa  Seung-Hee Kim
This study explores the feasibility of implementing a database (DB) access control solution by analyzing the PostgreSQL Front-End/Back-End (FEBE) protocol. To this end, algorithms were designed to extract session metadata and SQL text from startup, query, and parse packets. These algorithms enable control over user and database access, as well as SQL commands, including DML, DDL, and DCL types. The extraction process was validated using Apache JMeter, simulating concurrent SQL sessions. Results showed 100% accuracy across 100 session tests and 200 SQL executions.Unlike conventional PostgreSQL DB security tools focused on auditing and logging, the proposed method emphasizes real-time detection and protocol-level analysis. It allows for session-aware filtering, SQL pattern-based control, and fine-grained object-level restriction. The study demonstrates that protocol-layer information can serve as a viable foundation for developing PostgreSQL security solutions. Furthermore, this approach shows potential for extending to other open-source DBMSs, enhancing the scope and applicability of protocol-based security mechanisms.
The Transactions of the Korea Information Processing Society, Vol. 14, No. 8, pp. 595-607, Aug. 2025
https://doi.org/10.3745/TKIPS.2025.14.8.595
Database access control FEBE Protocol Analysis intrusion detection system (IDS) PostgreSQL Security
Persona-based One-shot MBTI Prompt Engineering
Yongjae Lee  Youngmin Ji  Yunsick Sung
Recently, large-scale language models are utilized in the diverse kinds of domains. However, It is still required to overcome the limitations of conventional generic and neutral response styles of a conversational agent as well as large-scale language models that are biased toward specific MBTI types. This paper proposes a dialogue generation technique that effectively reflects a conversational agent’s MBTI-based personality traits to provide personalized responses to users. This paper designs a system that conveys an agent’s personality traits with only few prompt examples by utilizing one-shot prompt engineering technique. The proposed system is composed of learning and inference phases; during the learning phase, the agent’s dialogues are evaluated and calibrated to ensure consistency with the agent’s persona and MBTI characteristics. During the inference phase, the optimized MBTI prompts are selectively applied to enable the agent to maintain rich contextual understanding and consistent personality. This approach presents the potential to deliver more natural and personalized interactive experiences in various fields, including virtual assistants, educational tutors, and entertainment characters
The Transactions of the Korea Information Processing Society, Vol. 14, No. 8, pp. 608-616, Aug. 2025
https://doi.org/10.3745/TKIPS.2025.14.8.608
One-shot Prompt MBTI Persona prompt engineering
A Deep Learning-Based Real-Time Vehicle Stop Prediction Method Using Velocity Trend Regression
Seung Hwan Cheon  Soo Hyung Kim
Conventional enforcement of illegal parking operates on a post-detection basis, where violations are confirmed only after the driver has exited the vehicle. This reactive approach limits the ability to maintain smooth traffic flow and ensure pedestrian safety. Addressing this issue through an IT-based system presents a significant challenge, as it requires high accuracy in determining whether a vehicle has stopped for parking and when to classify it as being in a ‘parked state.’ Moreover, various stop-related regulations, such as temporary halts before right turns, are stipulated in traffic laws, highlighting the broader applicability and importance of early stop prediction technology in both accident prevention and driver accountability. This study proposes a speed trend regression method to predict vehicle stoppage at an early stage. The system integrates YOLOv8 for object detection and ByteTrack for object tracking. By analyzing vehicle speed trends, it classifies vehicles into stopped, moving, and near-stopped categories. Experimental results, based on five test videos, show a Prediction Detection Performance (PDP) rate of 96% and a precision of 0.875, demonstrating the effectiveness and practical applicability of the proposed approach.
The Transactions of the Korea Information Processing Society, Vol. 14, No. 8, pp. 617-626, Aug. 2025
https://doi.org/10.3745/TKIPS.2025.14.8.617
Illegal parking Backstreets object tracking Vehicle stopping stop prediction
Design of Depth Adjustable Neural Networks for Vision Tasks
Kang Woochul
In this paper, we propose the architecture and training method of a depth adjustable deep neural network that enables on-the-fly selection of different inference accuracy-efficiency trade-offs using a single trained model. The proposed depth adjustable network splits each residual stage into two parts: one part learns hierarchical representations, while the other refines the learned features using a self-distillation technique. This training strategy allows the model to jointly learn multiple embedded subnetworks of varying depths with at most twice the training time of a single network. The proposed depth adjustable model is applied to image classification and object detection tasks, demonstrating the ability to adjust the trade-off between accuracy and efficiency without sacrificing maximum accuracy
The Transactions of the Korea Information Processing Society, Vol. 14, No. 8, pp. 627-632, Aug. 2025
https://doi.org/10.3745/TKIPS.2025.14.8.627
Deep Learning Adaptability Efficiency depth adjustment
A Study on Optimization Methods Based on Generative Flow Networks for Airport Resource Allocation Problems
Kim Jun Kyu  Ahn Kwang Mo
With the continuous growth in global air traffic demand, the operational efficiency of airports has become a critical factor in the competitiveness of the aviation industry. Among various operational challenges, the Gate Assignment Problem (GAP) is a combinatorial optimization problem characterized by complex resource constraints and real-time requirements. Existing approaches have shown limitations in terms of solution diversity and computational efficiency. In this study, GAP is mathematically modeled as a Mixed Integer Linear Programming (MILP) formulation, incorporating practical operational constraints such as time windows, gate capacity, and airline preferences. A novel optimization method based on Generative Flow Networks (GFlowNet) is proposed to learn reward-proportional probability distributions and stochastically generate structurally diverse, high-reward solutions. Using real-world data from Incheon International Airport, the proposed GFlowNet-based approach was compared under the same experimental conditions with an A3C (Asynchronous Advantage Actor-Critic)-based reinforcement learning method. The results showed that the proposed method achieved a 100% gate utilization rate and an average execution time of 3.16 seconds, improving resource utilization by 5.4% and reducing runtime by 41% compared to A3C. Furthermore, it achieved over 95% constraint satisfaction and a diversity score of 1.0, demonstrating its effectiveness in exploring non-redundant, high-quality solutions. This study empirically demonstrates that combining MILP-based mathematical modeling with generative reinforcement learning can effectively contribute to real-time airport operations and intelligent resource allocation.
The Transactions of the Korea Information Processing Society, Vol. 14, No. 8, pp. 633-642, Aug. 2025
https://doi.org/10.3745/TKIPS.2025.14.8.633
Airport operations Gate assignment combinatorial optimization Generative Flow Network MILP Reinforcement Learning
Performance Analysis of Personality Recognition on a Korean-Based Personality Recognition Dataset
Xu Qiu  Mira Lee  Jeehyeong Kim  Bongjae Kim
As the demand for personalized services continues to grow, research on personality recognition using artificial intelligence (AI) based on the OCEAN model has been rapidly advancing. However, there is still a lack of Korean-language personality recognition datasets. To address this, the Korea Electronics Technology Institute (KETI) recruited participants and constructed a Korean-language personality recognition dataset. This study utilized the dataset to compare two experimental settings: one in which the data was split into training, validation, and test sets based on participant IDs, and another in which the video data was randomly split into these sets. Additionally, experiments were conducted using both full videos and videos containing only speech segments within each experimental setting. In the participant-based data partitioning experiments, the experiment using the full video of the third scenario (SCENE 3) achieved the best 1-MAE performance (0.9142). In contrast, in the random split experiments, the integrated experiment using the full video across all three scenarios achieved the highest 1-MAE performance (0.9931). These results suggest that data preprocessing and splitting methods significantly impact the performance of Korean-language personality recognition models. This study is meaningful in that it provides foundational insights for future multimodal fusion-based personality recognition research.
The Transactions of the Korea Information Processing Society, Vol. 14, No. 8, pp. 643-649, Aug. 2025
https://doi.org/10.3745/TKIPS.2025.14.8.643
Artificial intelligence OCEAN Personality Recognition Transformer

Forms

Search
(IN TITLE, AUTHOR, ABSTRACT,KEYWORDS)

Advanced Search ▼

POPULAR KEYWORDS
(TOP 10 KEYWORDS)

Recent Publications
(LAST 3 YEARS)

Old Journals

Indexing

All publications of TKIPS is indexed in DOI, EBSCO, Google Scholar, Crossref, and CrossCheck.

TKIPS is also selected as the Journal for Accreditation by NRF (National Research Foundation of Korea).

Related Journals