A5062022079- Dental Radiography and Imaging
- Video Surveillance and Tracking Methods
- Dental Research and COVID-19
- Adversarial Robustness in Machine Learning
- Anomaly Detection Techniques and Applications
- Human Pose and Action Recognition
- Radiomics and Machine Learning in Medical Imaging
- Dental Health and Care Utilization
- COVID-19 diagnosis using AI
- Robotics and Sensor-Based Localization
- Forensic Toxicology and Drug Analysis
- Hand Gesture Recognition Systems
- Medical Imaging and Analysis
- Advanced X-ray and CT Imaging
- Advanced Image and Video Retrieval Techniques
- Hearing Impairment and Communication
University of Phayao
2013-2024
Shanghai Jiao Tong University
2022-2024
Chiang Mai University
2010
Abstract Currently, state-of-the-art object-tracking algorithms are facing a severe threat from adversarial attacks, which can significantly undermine their performance. In this research, we introduce MUNet, novel defensive model designed for visual tracking. This is capable of generating images that effectively counter attacks while maintaining low computational overhead. To achieve this, experiment with various configurations MUNet models, finding even minimal three-layer setup improves...
Abstract Currently, state-of-the-art object-tracking algorithms are facing a severe threat from adversarial attacks, which can significantly undermine their performance. In this work, we present the first defensive model for visual tracking, called MU-Net. The proposed generates denoised images that successfully defend against attacks while requiring minimal computing resources. To achieve this, MU-Net models explored in five different models, with smallest pair layers being 3. Each is...
The scale invariant feature transform (SIFT) has been used widely as a tool in object recognition. However, when there are several keyframes for one the training database, number of keypoint descriptors that might be huge. matching process test to done on all keypoints hence, amount time is Since database from same object, must some similar. In this paper we incorporate SIFT with Hard C-Means (HCM) algorithm group and then utilize prototypes instead. We implement three data sets, i.e.,...
Machine learning can categorize various algorithms, for example, classification model, object detection, segmentation and visual tracking. The adversarial attack comes up to addresses the generalization weakness which learns anticipate deep neural network by causing model malfunction. Although models effectively suppress performance of there are narrow attacking in tracking because is different from model. In this paper, we propose a feature-based that mildly generates perturbation template...