Qianqi Lu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Qianqi Lu has a total of 15 co-authors. Here are the top co-authors by number of shared publications:
- Xidao Luan
College of Computer Science and Engineering, Changsha University, Changsha, China
4 shared publications - Jing Zhang
College of Systems Engineering, National University of Defense Technology, Changsha, China
3 shared publications - Shiwei Zou
College of Systems Engineering, National University of Defense Technology, Changsha, China
3 shared publications - Yanming Guo
National University of Defense Technology, Changsha, Hunan, China
3 shared publications - Yuxiang Xie
College of Systems Engineering, National University of Defense Technology, Changsha, China
3 shared publications - Jing Zhang
College of Systems Engineering, National University of Defence Technology, Changsha 410000, China
2 shared publications
Latest Publications
- 16 January 2026
- Published by Springer Nature in International Journal of Multimedia Information Retrieval
- Vol. 15 (1) , 6
TFANet: three-stage image-text feature alignment network for robust referring image segmentation
- 16 January 2026
- Published by Springer Nature in International Journal of Multimedia Information Retrieval
- Vol. 15 (1) , 6
TFANet: three-stage image-text feature alignment network for robust referring image segmentation
- Open Access
- 8 November 2025
- Published by MDPI AG in Remote Sensing
- Vol. 17 (22) , 3675
CLIP-Driven with Dynamic Feature Selection and Alignment Network for Referring Remote Sensing Image Segmentation
Referring Remote Sensing Image Segmentation (RRSIS) aims to accurately locate and segment target objects in high-resolution aerial imagery based on natural language descriptions. Most existing approaches either directly modify Referring Image Segmentation (RIS) frameworks originally designed for natural images or employ image-based foundation models such as SAM to improve segmentation accuracy. However, current RRSIS models still face substantial challenges due to the domain gap between remote sensing and natural images, including large-scale variations, arbitrary object rotations, and complex spatial–linguistic relationships. Consequently, such transfers often lead to weak cross-modal interaction, inaccurate semantic alignment, and reduced localization precision, particularly for small or rotated objects. In addition, approaches that rely on multi-stage alignment pipelines, redundant high-level feature fusion, or the incorporation of large foundation models generally incur substantial computational overhead and training inefficiency, especially when dealing with complex referring expressions in high-resolution remote sensing imagery. To address these challenges, we propose CD2FSAN, a CLIP-driven dynamic feature selection and alignment network that establishes a unified framework for fine-grained cross-modal understanding in remote sensing imagery. This network first follows the principle of maximizing cross-modal information to dynamically select the visual representations most semantically aligned with the language from CLIP’s hierarchical features, thereby strengthening cross-modal correspondence under image domain shifts. It then performs adaptive multi-scale aggregation and alignment to integrate linguistic cues into spatially diverse visual contexts, enabling precise feature fusion across varying object scales. Finally, a dynamic rotation correction decoder with differentiable affine transformation was designed to refine segmentation by compensating for orientation diversity and geometric distortions. Extensive experiments verify that CD2FSAN consistently outperforms existing methods in segmentation accuracy, validating the effectiveness of its core components while maintaining competitive computational efficiency. These results demonstrate the framework’s strong capability to bridge the cross-modal gap between language and remote sensing imagery, highlighting its potential for advancing semantic understanding in vision–language remote sensing tasks.
- 8 November 2025
- Published by MDPI AG in Remote Sensing
- Vol. 17 (22) , 3675
Open AccessCLIP-Driven with Dynamic Feature Selection and Alignment Network for Referring Remote Sensing Image Segmentation
Referring Remote Sensing Image Segmentation (RRSIS) aims to accurately locate and segment target objects in high-resolution aerial imagery based on natural language descriptions. Most existing approaches either directly modify Referring Image Segmentation (RIS) frameworks originally designed for natural images or employ image-based foundation models such as SAM to improve segmentation accuracy. However, current RRSIS models still face substantial challenges due to the domain gap between remote sensing and natural images, including large-scale variations, arbitrary object rotations, and complex spatial–linguistic relationships. Consequently, such transfers often lead to weak cross-modal interaction, inaccurate semantic alignment, and reduced localization precision, particularly for small or rotated objects. In addition, approaches that rely on multi-stage alignment pipelines, redundant high-level feature fusion, or the incorporation of large foundation models generally incur substantial computational overhead and training inefficiency, especially when dealing with complex referring expressions in high-resolution remote sensing imagery. To address these challenges, we propose CD2FSAN, a CLIP-driven dynamic feature selection and alignment network that establishes a unified framework for fine-grained cross-modal understanding in remote sensing imagery. This network first follows the principle of maximizing cross-modal information to dynamically select the visual representations most semantically aligned with the language from CLIP’s hierarchical features, thereby strengthening cross-modal correspondence under image domain shifts. It then performs adaptive multi-scale aggregation and alignment to integrate linguistic cues into spatially diverse visual contexts, enabling precise feature fusion across varying object scales. Finally, a dynamic rotation correction decoder with differentiable affine transformation was designed to refine segmentation by compensating for orientation diversity and geometric distortions. Extensive experiments verify that CD2FSAN consistently outperforms existing methods in segmentation accuracy, validating the effectiveness of its core components while maintaining competitive computational efficiency. These results demonstrate the framework’s strong capability to bridge the cross-modal gap between language and remote sensing imagery, highlighting its potential for advancing semantic understanding in vision–language remote sensing tasks.
- Open Access
- 17 September 2025
- Published by MDPI AG in Preprints.org
CLIP-Driven with Dynamic Feature Selection and Alignment Network for Referring Remote Sensing Image Segmentation
Referring Remote Sensing Image Segmentation (RRSIS) aims to accurately locate and segment target objects in high-resolution aerial imagery based on natural language descriptions. The current RRSIS model faces numerous gaps due to the significant differences between remote sensing images and natural images, including scale variations, object rotation, and the difficulty of matching complex linguistic queries with spatially variable targets. Existing methods often rely on high-level semantic features or multi-stage cross-modal alignment, resulting in long training times and inefficiencies with complex queries. In this context, we propose CLIP-Driven with Dynamic Feature Selection and Alignment Network(CD2FSAN) , a novel framework that includes information-theoretic feature selection, adaptive multi-scale aggregation and alignment , and dynamic rotation correction decoder to better align remote sensing visual features with textual descriptions. Specially CD2FSAN dynamically selects visual features that best match the language description based on the principle of maximizing cross-modal information, alleviating domain shift from the pretraining of CLIP on natural images, and integrates language information during encoding. The framework also incorporates a multi-scale feature aggregation and alignment mechanism, ensuring precise cross-modal alignment, particularly for small targets. Additionally, CD2FSAN introduces a differentiable affine transformation-based dynamic rotation correction mechanism, enabling the network to adaptively adjust object orientations, improving segmentation accuracy. Experiments on three standard datasets, RefSegRS, RRSIS-D, and RISBench, demonstrate CD2FSAN’s superior performance in terms of oIoU, mIoU, and precision. Ablation studies and qualitative visualizations validate the efficacy of each module, confirming the framework’s robustness in handling spatial variations, rotation, and cross-modal alignment, significantly reducing the cross modality gap in CLIP-based single-stage RRSIS tasks.
- 17 September 2025
- Published by MDPI AG in Preprints.org
Open AccessCLIP-Driven with Dynamic Feature Selection and Alignment Network for Referring Remote Sensing Image Segmentation
Referring Remote Sensing Image Segmentation (RRSIS) aims to accurately locate and segment target objects in high-resolution aerial imagery based on natural language descriptions. The current RRSIS model faces numerous gaps due to the significant differences between remote sensing images and natural images, including scale variations, object rotation, and the difficulty of matching complex linguistic queries with spatially variable targets. Existing methods often rely on high-level semantic features or multi-stage cross-modal alignment, resulting in long training times and inefficiencies with complex queries. In this context, we propose CLIP-Driven with Dynamic Feature Selection and Alignment Network(CD2FSAN) , a novel framework that includes information-theoretic feature selection, adaptive multi-scale aggregation and alignment , and dynamic rotation correction decoder to better align remote sensing visual features with textual descriptions. Specially CD2FSAN dynamically selects visual features that best match the language description based on the principle of maximizing cross-modal information, alleviating domain shift from the pretraining of CLIP on natural images, and integrates language information during encoding. The framework also incorporates a multi-scale feature aggregation and alignment mechanism, ensuring precise cross-modal alignment, particularly for small targets. Additionally, CD2FSAN introduces a differentiable affine transformation-based dynamic rotation correction mechanism, enabling the network to adaptively adjust object orientations, improving segmentation accuracy. Experiments on three standard datasets, RefSegRS, RRSIS-D, and RISBench, demonstrate CD2FSAN’s superior performance in terms of oIoU, mIoU, and precision. Ablation studies and qualitative visualizations validate the efficacy of each module, confirming the framework’s robustness in handling spatial variations, rotation, and cross-modal alignment, significantly reducing the cross modality gap in CLIP-based single-stage RRSIS tasks.
- Open Access
TFANet: Three-Stage Image-Text Feature Alignment Network for Robust Referring Image Segmentation
Referring Image Segmentation (RIS) is a task that segments image regions based on language expressions, requiring fine-grained alignment between two modalities. However, existing methods often struggle with multimodal misalignment and language semantic loss, especially in complex scenes containing multiple visually similar objects, where uniquely described targets are frequently mislocalized or incompletely segmented. To tackle these challenges, this paper proposes TFANet, a Three-stage Image-Text Feature Alignment Network that systematically enhances multimodal alignment through a hierarchical framework comprising three stages: Knowledge Plus Stage (KPS), Knowledge Fusion Stage (KFS), and Knowledge Intensification Stage (KIS). In the first stage, we design the Multiscale Linear Cross-Attention Module (MLAM), which facilitates bidirectional semantic exchange between visual features and textual representations across multiple scales. This establishes rich and efficient alignment between image regions and different granularities of linguistic descriptions. Subsequently, the KFS further strengthens feature alignment through the Cross-modal Feature Scanning Module (CFSM), which applies multimodal selective scanning to capture long-range dependencies and construct a unified multimodal representation. This is essential for modeling long-range cross-modal dependencies and enhancing alignment accuracy in complex scenes. Finally, in the KIS, we propose the Word-level Linguistic Feature-guided Semantic Deepening Module (WFDM) to compensate for semantic degradation introduced in earlier stages.
Open AccessTFANet: Three-Stage Image-Text Feature Alignment Network for Robust Referring Image Segmentation
Referring Image Segmentation (RIS) is a task that segments image regions based on language expressions, requiring fine-grained alignment between two modalities. However, existing methods often struggle with multimodal misalignment and language semantic loss, especially in complex scenes containing multiple visually similar objects, where uniquely described targets are frequently mislocalized or incompletely segmented. To tackle these challenges, this paper proposes TFANet, a Three-stage Image-Text Feature Alignment Network that systematically enhances multimodal alignment through a hierarchical framework comprising three stages: Knowledge Plus Stage (KPS), Knowledge Fusion Stage (KFS), and Knowledge Intensification Stage (KIS). In the first stage, we design the Multiscale Linear Cross-Attention Module (MLAM), which facilitates bidirectional semantic exchange between visual features and textual representations across multiple scales. This establishes rich and efficient alignment between image regions and different granularities of linguistic descriptions. Subsequently, the KFS further strengthens feature alignment through the Cross-modal Feature Scanning Module (CFSM), which applies multimodal selective scanning to capture long-range dependencies and construct a unified multimodal representation. This is essential for modeling long-range cross-modal dependencies and enhancing alignment accuracy in complex scenes. Finally, in the KIS, we propose the Word-level Linguistic Feature-guided Semantic Deepening Module (WFDM) to compensate for semantic degradation introduced in earlier stages.
- Open Access
Video-based Generalized Category Discovery via Memory-Guided Consistency-Aware Contrastive Learning
Generalized Category Discovery (GCD) is an emerging and challenging open-world problem that has garnered increasing attention in recent years. Most existing GCD methods focus on discovering categories in static images. However, relying solely on static visual content is often insufficient to reliably discover novel categories. To bridge this gap, we extend the GCD problem to the video domain and introduce a new setting, termed Video-GCD. Thus, effectively integrating multi-perspective information across time is crucial for accurate Video-GCD. To tackle this challenge, we propose a novel Memory-guided Consistency-aware Contrastive Learning (MCCL) framework, which explicitly captures temporal-spatial cues and incorporates them into contrastive learning through a consistency-guided voting mechanism. MCCL consists of two core components: Consistency-Aware Contrastive Learning(CACL) and Memory-Guided Representation Enhancement (MGRE). CACL exploits multiperspective temporal features to estimate consistency scores between unlabeled instances, which are then used to weight the contrastive loss accordingly. MGRE introduces a dual-level memory buffer that maintains both feature-level and logit-level representations, providing global context to enhance intra-class compactness and inter-class separability. This in turn refines the consistency estimation in CACL, forming a mutually reinforcing feedback loop between representation learning and consistency modeling. To facilitate a comprehensive evaluation, we construct a new and challenging Video-GCD benchmark, which includes action recognition and bird classification video datasets. Extensive experiments demonstrate that our method significantly outperforms competitive GCD approaches adapted from image-based settings, highlighting the importance of temporal information for discovering novel categories in videos. The code will be publicly available.
Open AccessVideo-based Generalized Category Discovery via Memory-Guided Consistency-Aware Contrastive Learning
Generalized Category Discovery (GCD) is an emerging and challenging open-world problem that has garnered increasing attention in recent years. Most existing GCD methods focus on discovering categories in static images. However, relying solely on static visual content is often insufficient to reliably discover novel categories. To bridge this gap, we extend the GCD problem to the video domain and introduce a new setting, termed Video-GCD. Thus, effectively integrating multi-perspective information across time is crucial for accurate Video-GCD. To tackle this challenge, we propose a novel Memory-guided Consistency-aware Contrastive Learning (MCCL) framework, which explicitly captures temporal-spatial cues and incorporates them into contrastive learning through a consistency-guided voting mechanism. MCCL consists of two core components: Consistency-Aware Contrastive Learning(CACL) and Memory-Guided Representation Enhancement (MGRE). CACL exploits multiperspective temporal features to estimate consistency scores between unlabeled instances, which are then used to weight the contrastive loss accordingly. MGRE introduces a dual-level memory buffer that maintains both feature-level and logit-level representations, providing global context to enhance intra-class compactness and inter-class separability. This in turn refines the consistency estimation in CACL, forming a mutually reinforcing feedback loop between representation learning and consistency modeling. To facilitate a comprehensive evaluation, we construct a new and challenging Video-GCD benchmark, which includes action recognition and bird classification video datasets. Extensive experiments demonstrate that our method significantly outperforms competitive GCD approaches adapted from image-based settings, highlighting the importance of temporal information for discovering novel categories in videos. The code will be publicly available.