Pha Nguyen

Dr. Pha Nguyen

orcid

Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
30 March 2024 in Pattern Recognition
Verified
Journal Article
Closed access
PNPha Nguyen
Pha Nguyen
KQKha Gia Quach
Kha Gia Quach
CDChi Nhan Duong
Chi Nhan Duong
SPSon Lam Phung
Son Lam Phung
NLNgan Le
Ngan Le
KLKhoa Luu
Khoa Luu

13 February 2025 in IEEE Access
Verified
Journal Article
Open access
NCNaga Venkata Sai Raviteja Chappa
Naga Venkata Sai Raviteja Chappa
PNPha Nguyen
Pha Nguyen
ANAlexander H. Nelson
Alexander H. Nelson
Han-Seok SeoHan-Seok Seo
Han-Seok Seo
XLXin Li
Xin Li
KLKhoa Luu
Khoa Luu
Social group activity recognition is crucial for various applications including surveillance, human-robot interaction, and behavioral analysis. Current approaches often require extensive manual annotations and rely heavily on pre-trained detectors, limiting their practical applications. Additionally, existing methods struggle to effectively model long-term spatiotemporal relationships in group activities. This paper introduces a novel approach to Social Group Activity Recognition (SoGAR) using Self-supervised Transformers network that can effectively utilize unlabeled video data. To extract spatio-temporal information, we create local and global views with varying frame rates. Our self-supervised objective ensures that features extracted from contrasting views of the same video are consistent across spatio-temporal domains. Our proposed approach efficiently uses transformer-based encoders to alleviate the weakly supervised setting of group activity recognition. By leveraging the benefits of transformer models, our approach can model long-term relationships along spatio-temporal dimensions. Our proposed SoGAR method achieves state-of-the-art results on three group activity recognition benchmarks, namely JRDB-PAR, NBA, and Volleyball datasets, surpassing the current state-of-the-art in terms of F1-score, MCA, and MPCA metrics.

26 September 2024
Verified
Book Chapter
Closed access
PDF file
KQKha Gia Quach
Kha Gia Quach
PNPha Nguyen
Pha Nguyen
CDChi Nhan Duong
Chi Nhan Duong
TBTien Dai Bui
Tien Dai Bui
KLKhoa Luu
Khoa Luu

24 May 2024 in Sensors
Verified
Journal Article
Open access
PDF file
NCNaga Venkata Sai Raviteja Chappa
Naga Venkata Sai Raviteja Chappa
PNPha Nguyen
Pha Nguyen
HLHoang Ngan Le
Hoang Ngan Le
Page D DobbsPage D Dobbs
Page D Dobbs
KLKhoa Luu
Khoa Luu
Group-activity scene graph (GASG) generation is a challenging task in computer vision, aiming to anticipate and describe relationships between subjects and objects in video sequences. Traditional video scene graph generation (VidSGG) methods focus on retrospective analysis, limiting their predictive capabilities. To enrich the scene-understanding capabilities, we introduced a GASG dataset extending the JRDB dataset with nuanced annotations involving appearance, interaction, position, relationship, and situation attributes. This work also introduces an innovative approach, a Hierarchical Attention–Flow (HAtt-Flow) mechanism, rooted in flow network theory to enhance GASG performance. Flow–attention incorporates flow conservation principles, fostering competition for sources and allocation for sinks, effectively preventing the generation of trivial attention. Our proposed approach offers a unique perspective on attention mechanisms, where conventional “values” and “keys” are transformed into sources and sinks, respectively, creating a novel framework for attention-based models. Through extensive experiments, we demonstrate the effectiveness of our Hatt-Flow model and the superiority of our proposed flow–attention mechanism. This work represents a significant advancement in predictive video scene understanding, providing valuable insights and techniques for applications that require real-time relationship prediction in video data.

20 July 2024 in Machine Vision and Applications
Verified
Journal Article
Open access
PDF file
NCNaga Venkata Sai Raviteja Chappa
Naga Venkata Sai Raviteja Chappa
PNPha Nguyen
Pha Nguyen
Page D DobbsPage D Dobbs
Page D Dobbs
KLKhoa Luu
Khoa Luu
In the realm of computer vision, Group Activity Recognition (GAR) plays a vital role, finding applications in sports video analysis, surveillance, and social scene understanding. This paper introducesRecognizeEveryAction Everywhere All At Once (REACT), a novel architecture designed to model complex contextual relationships within videos. REACT leverages advanced transformer-based models for encoding intricate contextual relationships, enhancing understanding of group dynamics. Integrated Vision-Language Encoding facilitates efficient capture of spatiotemporal interactions and multi-modal information, enabling comprehensive scene understanding. The model’s precise action localization refines joint understanding of text and video data, enabling precise bounding box retrieval and enhancing semantic links between textual descriptions and visual reality. Actor-Specific Fusion strikes a balance between actor-specific details and contextual information, improving model specificity and robustness in recognizing group activities. Experimental results demonstrate REACT’s superiority over state-of-the-art GAR approaches, achieving higher accuracy in recognizing and understanding group activities across diverse datasets. This work significantly advances group activity recognition, offering a robust framework for nuanced scene comprehension.

1 October 2020
Verified
Conference Paper
Closed access
PNPha Nguyen
Pha Nguyen
STSon T. Tran
Son T. Tran
Object tracking is a fundamental domain in computer vision, especially multiple human tracking, which is an important task to solve many key problems, including traffic counting, behavior analysis of in-store customers. It is very difficult to handle these problems in heavily crowded scenes because of pedestrian occlusions. Recent research works focus on tracking-by-detection, which breaks tracking problem down into two steps: detect objects then associate them with groups by motion or features. Recent object detectors still do not have the accuracy-speed balance and do not work perfectly in a crowded area. In this paper, we propose a new method for human tracking, which is less depended on the instability of the object detector and inherits the success of Siamese Visual Tracking in recent years. While Siamese-family trackers work on a single object, we extend the problem by proposing a strategy that manages across multiple tracker and combines a human detector as a semi-supervisor for tracker location correction. Additionally, the detector-tracker associating strategy adapts the transformation of human poses and the acquisition of new pedestrians during the tracking process. Our method outperforms the state-of-the-art algorithms on our crowded scenes dataset.