Ploysuwan, Tuchsanai

Now showing 1 - 5 of 5

Some of the metrics are blocked by your
consent settings
Reducing the Counting Time of Colonies Using Image Processing Techniques
(2024-01-01)
Siangchin, Apichai
;
Ploysuwan, Tuchsanai
Microbial colony counting is a crucial process in microbiology laboratories and medical research, yet it is often time-consuming and labor-intensive, particularly in large-scale settings or when dealing with numerous samples. Incorporating image processing techniques can significantly reduce the time and error associated with manual colony counting. In this study, we developed and evaluated a deep learning model for automated E. coli colony counting using the YOLO (You Only Look Once) framework. The model achieved an mAP50 of 90.3%, demonstrating its high accuracy and potential for real-world application in laboratories of various sizes. This research not only streamlines the colony counting process but also allows laboratory personnel to allocate their time to other essential tasks.
Some of the metrics are blocked by your
consent settings
Dynamic System Linearization via Deep Relative Displacement Prediction for Robust IMU-WiFi Indoor Trajectory Estimation
(2025-01-01)
Chapha, Naphat
;
Ploysuwan, Tuchsanai
Standalone WiFi fingerprinting and Inertial Measurement Unit (IMU) sensors are unreliable for indoor positioning due to signal instability and cumulative drift error, respectively. We present a novel sensor fusion framework that overcomes these challenges by enhancing WiFi fingerprint quality while simplifying the fusion process. Our approach filters and ranks WiFi signals to create a stable input for a Long Short-Term Memory (LSTM), while a Convolutional Neural Network (CNN) is uniquely trained to predict relative displacement from IMU data. This latter step linearizes the system dynamics, enabling the use of a simple Linear Kalman Filter and avoiding the complexity of traditional non-linear filters. Combining these innovations, the integrated framework achieves a mean position error of 4.30 m, a 32.3% improvement over the best standalone model. This work demonstrates that our approach to WiFi selection and system linearization provides a robust and highly accurate solution for real-time indoor positioning.
Some of the metrics are blocked by your
consent settings
Code Smell Classification Using Graph Convolutional Network with Imbalanced Data and Model Integration
(2025-01-01)
Suphawimon, Phawinee
;
Ploysuwan, Tuchsanai
Code smell represents critical design anomalies that significantly impact software maintainability and quality. This paper presents a comprehensive framework using Graph Convolutional Networks (GCNs) integrated with traditional machine learning techniques. We systematically evaluated graph construction approaches, model integration methodologies, and data balancing strategies using nine real-world Python repositories labeled with PyExamine. Our methodology combines BERT embeddings with graph structural representations, implementing layer integration (Method I) and feature concatenation (Method II). Results show per-line graph construction outperforms global approaches, with SMOTE achieving 96.14% accuracy compared to 86.70% for imbalanced data. Including non-smelly code improves performance from 71% to 95%, demonstrating the importance of negative examples. Our ablation study shows explicit feature engineering achieves only 67% accuracy compared to 95% for end-to-end learning. The integrated GCN with Transformer using Method II achieved 95% accuracy and 89% F1-score, nearly matching CodeT5 (97% accuracy, 85% F1-score) while providing better interpretability.
Some of the metrics are blocked by your
consent settings
Fall Detection Approach Using Variational Autoencoders with Self-Attention Features
(2023-01-01)
Soontornnapar, Tomorn
;
Ploysuwan, Tuchsanai
In this paper, we propose an alternative method for fall detection using variational autoencoders (VAEs) with an attention mechanism on an existing dataset. The dataset consists of 6 different fall cases from 21 people. For effective fall detection, we introduce the use of the magnitude of the acceleration vector (MAV) of wearable gyroscope data and apply fast-Fourier transform (FFT) to create new features. These FFT features are then passed through attention modules with self-combination to form attention features. Our experimental results show that the VAE with self-attention features achieved an average accuracy of 90.7% and an F1 score of 93.8% in fall detection, demonstrating the effectiveness of the proposed method in utilizing gyroscope sensors for fall detection in the context of threshold criteria.
Some of the metrics are blocked by your
consent settings
Zero-/Few-Shot Anomaly Classification for Transistor Using Multimodal CLIP Retrieval Augmented
(2024-01-01)
Praneenatthavee, Natdhanai
;
Ploysuwan, Tuchsanai
Detecting anomalies in transistors is a challenging task due to the intricate features distinguishing normal from abnormal components. This paper introduces a zero-shot and few-shot anomaly classification framework for transistors using a retrieval-augmented multimodal approach. We focus on two critical parts of the transistor: the Component BOX and the Metal Legs. Utilizing Florence2 for prompt-based bounding box generation and Segment Anything Model2 (SAM2) for segmentation, we create precise masks for each part. Embeddings generated through Contrastive Language-Image Pre-training (CLIP) are employed to classify each component effectively. For the few-shot learning scenario, we implement Retrieval-Augmented Generation (RAG) to simulate learning from both images and textual data, enhancing the anomaly classification performance. Our zero-shot model achieved an fl score of 70.5, while the few-shot model attained an improved fl score of 77.0, demonstrating the efficacy of our approach in transistor anomaly detection.

Ploysuwan, Tuchsanai

Filters

Settings

Sort By

Results per page