Yaqian Zhao
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Yaqian Zhao has a total of 74 co-authors. Here are the top co-authors by number of shared publications:
- Rengang Li
Southeast University
7 shared publications - Ruyang Li
Inspur (Beijing) Electronic Information Industry Co., Ltd
5 shared publications - Xin Zhang
College of Intelligence and Computing Tianjin University Tianjin China
4 shared publications - Jinzhe Jiang
IEIT SYSTEMS (Beijing) Co., Ltd
3 shared publications - Qi Deng
Inspur (Beijing) Electronic Information Industry Co., Ltd
3 shared publications - Zhenhua Guo
IEIT SYSTEMS Co., Ltd., Jinan, China
3 shared publications
Latest Publications
- 26 September 2026
- Published by Springer Nature
- p. 537-550
FLESC: Flexible Stripeless Erasure Coding Scheme for In-Memory Storage Systems
- 26 September 2026
- Published by Springer Nature
- p. 537-550
FLESC: Flexible Stripeless Erasure Coding Scheme for In-Memory Storage Systems
- 9 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. PP (15209210) , 1-13
Learning Semantic-level Multi-modal Alignment Transformer for RGB-D Co-salient Object Detection
The goal of RGB-D Co-Salient Object Detection (CoSOD) is to integrate both depth and RGB information to identify common and salient objects across a set of images. However, the existing leading research paradigm often fuses raw depth and RGB cues indiscriminately via simple concatenation or weighted summation, neglecting a fundamental limitation: the low inter object contrast in depth maps often causes misalignment between RGB-salient regions and depth-highlighted areas, resulting in model performance degradation. To address this discrepancy, we propose a Semantic-level Multi-modal Alignment transFormer (SMAFormer) that learns to establish precise correspondence for the co-salient objects across RGB and depth modalities. The SMAFormer mainly incorporates two key components: a Learnable Depth-calibration Module (LDM) and a cross-modal transformer (CMT). In the LDM, a Semantic-level Cluster center Extractor (SCE) is first designed to generate semantic level masks, which are then converted to semantic-level cluster centers via the Segment Anything Model (SAM). Afterwards, a Recurrent Deep Clustering Module (RDCM) iteratively projects these semantic representations onto the depth features through a learnable clustering strategy, accurately calibrating the depth features with rich semantic alignment information between two modalities. Next, in CMT, cross-modal shared representations between RGB and calibrated depth features are extracted and mixed through interaction-attention mechanism. Extensive evaluations demonstrate that SMAFormer establishes new state-of the-art performance across three benchmarks including RGB-D CoSal150, RGB-D CoSeg183, and RGB-D CoSal1k.
- 9 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. PP (15209210) , 1-13
Learning Semantic-level Multi-modal Alignment Transformer for RGB-D Co-salient Object Detection
The goal of RGB-D Co-Salient Object Detection (CoSOD) is to integrate both depth and RGB information to identify common and salient objects across a set of images. However, the existing leading research paradigm often fuses raw depth and RGB cues indiscriminately via simple concatenation or weighted summation, neglecting a fundamental limitation: the low inter object contrast in depth maps often causes misalignment between RGB-salient regions and depth-highlighted areas, resulting in model performance degradation. To address this discrepancy, we propose a Semantic-level Multi-modal Alignment transFormer (SMAFormer) that learns to establish precise correspondence for the co-salient objects across RGB and depth modalities. The SMAFormer mainly incorporates two key components: a Learnable Depth-calibration Module (LDM) and a cross-modal transformer (CMT). In the LDM, a Semantic-level Cluster center Extractor (SCE) is first designed to generate semantic level masks, which are then converted to semantic-level cluster centers via the Segment Anything Model (SAM). Afterwards, a Recurrent Deep Clustering Module (RDCM) iteratively projects these semantic representations onto the depth features through a learnable clustering strategy, accurately calibrating the depth features with rich semantic alignment information between two modalities. Next, in CMT, cross-modal shared representations between RGB and calibrated depth features are extracted and mixed through interaction-attention mechanism. Extensive evaluations demonstrate that SMAFormer establishes new state-of the-art performance across three benchmarks including RGB-D CoSal150, RGB-D CoSeg183, and RGB-D CoSal1k.
- Open Access
First Place Solution to the ECCV 2024 ROAD++ Challenge @ ROAD++ Spatiotemporal Agent Detection 2024
This report presents our team's solutions for the Track 1 of the 2024 ECCV ROAD++ Challenge. The task of Track 1 is spatiotemporal agent detection, which aims to construct an "agent tube" for road agents in consecutive video frames. Our solutions focus on the challenges in this task, including extreme-size objects, low-light scenarios, class imbalance, and fine-grained classification. Firstly, the extreme-size object detection heads are introduced to improve the detection performance of large and small objects. Secondly, we design a dual-stream detection model with a low-light enhancement stream to improve the performance of spatiotemporal agent detection in low-light scenes, and the feature fusion module to integrate features from different branches. Subsequently, we develop a multi-branch detection framework to mitigate the issues of class imbalance and fine-grained classification, and we design a pre-training and fine-tuning approach to optimize the above multi-branch framework. Besides, we employ some common data augmentation techniques, and improve the loss function and upsampling operation. We rank first in the test set of Track 1 for the ROAD++ Challenge 2024, and achieve 30.82% average video-mAP.
Open AccessFirst Place Solution to the ECCV 2024 ROAD++ Challenge @ ROAD++ Spatiotemporal Agent Detection 2024
This report presents our team's solutions for the Track 1 of the 2024 ECCV ROAD++ Challenge. The task of Track 1 is spatiotemporal agent detection, which aims to construct an "agent tube" for road agents in consecutive video frames. Our solutions focus on the challenges in this task, including extreme-size objects, low-light scenarios, class imbalance, and fine-grained classification. Firstly, the extreme-size object detection heads are introduced to improve the detection performance of large and small objects. Secondly, we design a dual-stream detection model with a low-light enhancement stream to improve the performance of spatiotemporal agent detection in low-light scenes, and the feature fusion module to integrate features from different branches. Subsequently, we develop a multi-branch detection framework to mitigate the issues of class imbalance and fine-grained classification, and we design a pre-training and fine-tuning approach to optimize the above multi-branch framework. Besides, we employ some common data augmentation techniques, and improve the loss function and upsampling operation. We rank first in the test set of Track 1 for the ROAD++ Challenge 2024, and achieve 30.82% average video-mAP.
- 16 June 2024
- Published by Springer Nature in Circuits, Systems, and Signal Processing
- Vol. 43 (9) , 5487-5506
Hermite Expansion Technique for Model Reduction of Circuit Systems with Delay Components
- 16 June 2024
- Published by Springer Nature in Circuits, Systems, and Signal Processing
- Vol. 43 (9) , 5487-5506
Hermite Expansion Technique for Model Reduction of Circuit Systems with Delay Components
- 27 May 2024
- Published by Elsevier in Chaos, Solitons & Fractals
- Vol. 184
Recurrent dynamic message passing with loops for epidemics on networks
- 27 May 2024
- Published by Elsevier in Chaos, Solitons & Fractals
- Vol. 184
Recurrent dynamic message passing with loops for epidemics on networks
- Open Access
- 4 February 2024
- Published by Springer Nature in Multimedia Systems
- Vol. 30 (1) , 1-10
MVIndEmo: a dataset for micro video public-induced emotion prediction on social media
Distinct from the realm of perceived emotion research, induced emotion pertains to the emotional responses engendered within content consumers. This facet has garnered considerable attention and finds extensive application in the analysis of public social media. However, the advent of micro videos presents unique challenges when attempting to discern the induced emotional patterns exhibited by content consumers, owing to their free-style representation and other factors. Consequently, we have put forth two novel tasks concerning the recognition of public-induced emotion on micro videos: emotion polarity and emotion classification. Additionally, we have introduced a accessible dataset specifically tailored for the analysis of public-induced emotion on micro videos. The data corpus has been meticulously collected from Tiktok, a burgeoning social media platform renowned for its trendsetting content. To construct the dataset, we have selected eight captivating topics that elicit vibrant social discussions. In devising our label generation strategy, we have employed an automated approach characterized by the fusion of multiple expert models. This strategy incorporates a confidence measure method that relies on three distinct models for effectively aggregating user comments. To accommodate adaptable benchmark configurations, we provide both binary classification labels and probability distribution labels. The dataset encompasses a vast collection of 7,153 labeled micro videos. We have undertaken an extensive statistical analysis of the dataset to provide a comprehensive overview composition. It is our earnest aspiration that this dataset will serve as a catalyst for pioneering research avenues in the analysis of emotional patterns and the understanding of multi-modal information.
- 4 February 2024
- Published by Springer Nature in Multimedia Systems
- Vol. 30 (1) , 1-10
Open AccessMVIndEmo: a dataset for micro video public-induced emotion prediction on social media
Distinct from the realm of perceived emotion research, induced emotion pertains to the emotional responses engendered within content consumers. This facet has garnered considerable attention and finds extensive application in the analysis of public social media. However, the advent of micro videos presents unique challenges when attempting to discern the induced emotional patterns exhibited by content consumers, owing to their free-style representation and other factors. Consequently, we have put forth two novel tasks concerning the recognition of public-induced emotion on micro videos: emotion polarity and emotion classification. Additionally, we have introduced a accessible dataset specifically tailored for the analysis of public-induced emotion on micro videos. The data corpus has been meticulously collected from Tiktok, a burgeoning social media platform renowned for its trendsetting content. To construct the dataset, we have selected eight captivating topics that elicit vibrant social discussions. In devising our label generation strategy, we have employed an automated approach characterized by the fusion of multiple expert models. This strategy incorporates a confidence measure method that relies on three distinct models for effectively aggregating user comments. To accommodate adaptable benchmark configurations, we provide both binary classification labels and probability distribution labels. The dataset encompasses a vast collection of 7,153 labeled micro videos. We have undertaken an extensive statistical analysis of the dataset to provide a comprehensive overview composition. It is our earnest aspiration that this dataset will serve as a catalyst for pioneering research avenues in the analysis of emotional patterns and the understanding of multi-modal information.
- Open Access
- 2 February 2024
- Published by Association for Computing Machinery (ACM)
- p. 94-99
Study on The Effect of Encoding Method in Quantum Machine Learning
Quantum machine learning algorithms use qubit encoding and quantum circuits to perform feature extraction and pattern recognition. However, the choice of data encoding methods can have a significant impact on the model’s performance. To evaluate the effect of different encoding methods in a systematic way, we propose two metrics: distribution distance and distribution radius. These metrics describe how the encoded data distribute in the Hilbert space. We show that there is a positive correlation between prediction accuracy and distribution distance, and a negative correlation between prediction accuracy and distribution radius, both theoretically and experimentally. Based on our findings, we suggest a comparative evaluation of data encoding methods for quantum machine learning, which can help improve the learning efficiency.
- 2 February 2024
- Published by Association for Computing Machinery (ACM)
- p. 94-99
Open AccessStudy on The Effect of Encoding Method in Quantum Machine Learning
Quantum machine learning algorithms use qubit encoding and quantum circuits to perform feature extraction and pattern recognition. However, the choice of data encoding methods can have a significant impact on the model’s performance. To evaluate the effect of different encoding methods in a systematic way, we propose two metrics: distribution distance and distribution radius. These metrics describe how the encoded data distribute in the Hilbert space. We show that there is a positive correlation between prediction accuracy and distribution distance, and a negative correlation between prediction accuracy and distribution radius, both theoretically and experimentally. Based on our findings, we suggest a comparative evaluation of data encoding methods for quantum machine learning, which can help improve the learning efficiency.
- Open Access
- 20 January 2024
- Published by Springer Nature in Discover Applied Sciences
- Vol. 6 (1) , 1-12
Enhancing and confining light in hybrid plasmonic nanowire-integrated V-groove silicon waveguides
In recent years, the field of dielectric-plasmonic photonics has made remarkable strides, leading to the successful development of various technologies. The realization of sophisticated optical circuits on a single platform has become increasingly viable. Here we propose and investigate a hybrid dielectric waveguide integrated with plasmonics. This hybrid optical waveguide comprises a copper nanowire situated in close proximity to a silicon V-groove channel, separated by a nanoscale gap. This configuration is particularly advantageous, as achieving precise alignment of the nanowire within the V-groove addresses a fundamental challenge in engineering a fully functional integrated component. Additionally, a silicon nitride film coats the V-groove. Utilizing finite element analysis, we conduct numerical simulations to analyze field properties and modal propagation at a specific wavelength of 1550 nm. Our simulations reveal that meticulous optimization of the nanowire and V-groove channel’s geometrical parameters enables effective tailoring of the hybrid mode. This optimization results in strong mode coupling between the dielectric waveguide mode and the surface plasmon, leading to substantial field enhancement, confinement, and extended propagation length. These waveguides also hold promise for sensing applications, facilitating the detection of sample variations and locations due to pronounced mode characteristics. The proposed hybrid approach demonstrates potential for integration into high-level photonic circuits and on-chip optical computing systems.
- 20 January 2024
- Published by Springer Nature in Discover Applied Sciences
- Vol. 6 (1) , 1-12
Open AccessEnhancing and confining light in hybrid plasmonic nanowire-integrated V-groove silicon waveguides
In recent years, the field of dielectric-plasmonic photonics has made remarkable strides, leading to the successful development of various technologies. The realization of sophisticated optical circuits on a single platform has become increasingly viable. Here we propose and investigate a hybrid dielectric waveguide integrated with plasmonics. This hybrid optical waveguide comprises a copper nanowire situated in close proximity to a silicon V-groove channel, separated by a nanoscale gap. This configuration is particularly advantageous, as achieving precise alignment of the nanowire within the V-groove addresses a fundamental challenge in engineering a fully functional integrated component. Additionally, a silicon nitride film coats the V-groove. Utilizing finite element analysis, we conduct numerical simulations to analyze field properties and modal propagation at a specific wavelength of 1550 nm. Our simulations reveal that meticulous optimization of the nanowire and V-groove channel’s geometrical parameters enables effective tailoring of the hybrid mode. This optimization results in strong mode coupling between the dielectric waveguide mode and the surface plasmon, leading to substantial field enhancement, confinement, and extended propagation length. These waveguides also hold promise for sensing applications, facilitating the detection of sample variations and locations due to pronounced mode characteristics. The proposed hybrid approach demonstrates potential for integration into high-level photonic circuits and on-chip optical computing systems.
- 5 December 2023
- Published by Springer Nature in Quantum Information Processing
- Vol. 22 (12) , 1-12
Strong generalization in quantum neural networks
- 5 December 2023
- Published by Springer Nature in Quantum Information Processing
- Vol. 22 (12) , 1-12
Strong generalization in quantum neural networks