arXiv subjects

Zhengjie Zhang

Publications and source records attributed to Zhengjie Zhang.

12 recordsLinked to original sources

arXiv2026

What is the Better Curriculum: Controller-Shaped Grasping Behavior for Contact Force-Sensitive Manipulation

How should a robot learn to manipulate objects so fragile that sub-Newton contact forces can cause irreversible damage? Existing visuo-tactile policy learning typically treats tactile sensing as an additional policy input. In direct-contact force-sensitive manipulation, however, the bottleneck can arise earlier, during data collection: manual gripper control is too delayed and coarse-grained to reliably maintain the narrow force range required for stable grasping. We therefore use a deterministic 25 Hz tactile reflex controller as a collection-time teacher, producing demonstrations with controller-shaped grasping behavior for tactile-free policy learning. On Action Chunking with Transformers (ACT), policies trained from reflex-shaped demonstrations recover the teacher's grasping profile and achieve 95% stable grasps on the nominal plastic-cup task, substantially outperforming visually screened manual demonstrations. The same intervention improves in-distribution stability on $π_{0.5}$ and shows a favorable exploratory trend on an unseen paper-cup variant. Under randomized external disturbance, however, the reflex-data $π_{0.5}$ policy still fails in 45% of policy-only trials, whereas a deployment-time reflex arbiter retains all grasps. These results reveal a new role for tactile feedback in force-sensitive manipulation: rather than integrating tactile into the policy, we use it as a collection-time teacher that shapes grasping behavior in demonstrations for policy learning, while disturbance rejection remains controller-dependent, revealing the boundary of tactile-free policy.

cs.RO↗

arXiv2026

RGSQ: Riemannian Geometry-Sensitive Quantization for Large Vision-Language Models

Large vision-language models (VLMs) can be efficiently deployed under stringent memory and latency constraints through post training quantization (PTQ). However, most PTQ methods are designed for unimodal large language models (LLMs). These methods treat quantization errors as isotropic perturbations under the Euclidean assumption, which provides weak guidance on directions most sensitive to quantization in VLMs. Consequently, directly adapting unimodal PTQ approaches or solely employing modality-specific scaling often leads to uneven bit-width distribution and inconsistent performance in low-bit settings. To address these challenges, we propose Riemannian Geometry-Sensitive Quantization (RGSQ), which formulates quantization as a reconstruction problem under a unified Fisher-Riemannian metric. RGSQ identifies modality-specific sensitive directions via Riemannian manifold mappings built from modality-partitioned empirical Fisher factors and fused into a modality-aware Kronecker-structured metric. We then apply geometry-aligned rotations to reorient the local tangent frame, steering low-bit perturbations toward loss-insensitive axes. Finally, we apply a whitening transformation that maps the Riemannian objective to an equivalent Euclidean form, enabling standard unimodal PTQ methods to evaluate multimodal quantization error under their original assumptions. Across an extensive and diverse set of mainstream VLM benchmarks, RGSQ achieves the highest accuracy and stability under extremely low-bit settings (W2A8 and W3A8). It outperforms VLM-aware baselines, such as MBQ and MQuant, by up to 5.9% and surpasses single-modality improvements by up to 8.6%.

cs.CV↗

arXiv2026

DeepInsight II: One Trace from Benchmark to Robot

Across a Physical AI stack, evaluation maturity is inversely aligned with deployment risk: foundation models enjoy mature, standardized harnesses, while the embodied layers on which deployment actually turns remain fragmented across benchmark-specific simulators, embodiments, and interfaces. The first DeepInsight report (v1) unified evaluation across this stack behind three abstractions---task, resource, and result---but its quantitative evidence centered on the foundation-model layer; navigation and manipulation (System 1) and whole-body control (System 0) remained simulation case studies, and physical execution was outside its empirical scope. DeepInsight II keeps that substrate fixed and quantifies the embodied half. First, it reproduces released-checkpoint references across two navigation and four manipulation benchmarks under their native protocols. Second, MotionBench places four released whole-body controllers under one workload and metric contract, then carries a qualified within-family cohort from parallel simulation to matched real-robot trials in which simulated and physical rollouts share a parent trace identity while retaining execution-domain-specific records, making the sim-to-real gap a native reduction rather than a reconciliation across toolchains. Third, a composed System 2--1--0 study extends trace localization into five evidence-grounded handoff labels, each mapped to a concrete repair action, with a measured repairability criterion and physical episodes testing the same attribution under hardware-observable state. The contribution is therefore not a new evaluation architecture, but empirical continuity from benchmark execution to matched robot evidence and repair-oriented diagnosis.

cs.AI↗

arXiv2026

ST-ColoNet: Spatio-Temporal Colon Segment Recognition via Hybrid Attention and Edge-Guided Feature Learning

Colo-segment recognition in colonoscopy videos is a key requirement for many downstream tasks, but existing automatic recognition methods only use colonoscopy images without fully exploiting the use of temporal information, leading to poor performance. Additionally, relevant public video-based datasets are in scarcity. To tackle this problem, we curate and release a labeled dataset specifically for the task of colo-segment recognition. In addition, we propose a two-stage deep learning-based framework, Colo-Segment Recognition via SpatioTemporal Network (ST-ColoNet), for the task of colo-segment recognition from colonoscopy videos which includes the Colorlaus module that uses metric learning to optimize edge-mediated spatial feature extraction, as well as the Full-Temp module which combines three self-attention patterns to better approximate full self-attention on long colonoscopy sequences and optimize temporal feature aggregation. Through extensive ablation experiments, we show that our framework is capable of achieving state-of-the-art performance on the task of colo-segment recognition, achieving an accuracy of 81.0% and F1-score of 70.7%, which is a tremendous improvement over state-of-the-art methods.

cs.CV↗

arXiv2025

Colo-ReID: Discriminative Representation Embedding with Meta-learning for Colonoscopic Polyp Re-Identification

Colonoscopic Polyp Re-Identification aims to match the same polyp from a large gallery with images from different views taken using different cameras and plays an important role in the prevention and treatment of colorectal cancer. However, traditional methods for object ReID directly adopting CNN models trained on the ImageNet dataset usually produce unsatisfactory retrieval performance on colonoscopic datasets due to the large domain gap. Additionally, these methods neglect to explore the potential of self-discrepancy among intra-class or inter-class relations in the colonoscopic polyp dataset, which remains an open research problem in the medical community. To solve this dilemma, we propose a simple but effective training method named Colo-ReID, which can help our model learn more general and discriminative knowledge based on the meta-learning strategy in scenarios with fewer samples. Based on this, a dynamic Meta-Learning Regulation mechanism called MLR is introduced to further boost the performance of polyp re-identification. Our experimental results show that Colo-ReID consistently outperforms second-best method in terms of mAP performance by +2.3% on polyp re-identification task. Our source code is also publicly available at https://github.com/JeremyXSC/Colo-ReID.

cs.CV↗

arXiv2025

Language-Enhanced Generative Modeling for Amyloid PET Synthesis from MRI and Blood Biomarkers

Background: Alzheimer's disease (AD) diagnosis heavily relies on amyloid-beta positron emission tomography (Abeta-PET), which is limited by high cost and limited accessibility. This study explores whether Abeta-PET spatial patterns can be predicted from blood-based biomarkers (BBMs) and MRI scans. Methods: We collected Abeta-PET images, T1-weighted MRI scans, and BBMs from 566 participants. A language-enhanced generative model, driven by a large language model (LLM) and multimodal information fusion, was developed to synthesize PET images. Synthesized images were evaluated for image quality, diagnostic consistency, and clinical applicability within a fully automated diagnostic pipeline. Findings: The synthetic PET images closely resemble real PET scans in both structural details (SSIM = 0.920 +/- 0.003) and regional patterns (Pearson's r = 0.955 +/- 0.007). Diagnostic outcomes using synthetic PET show high agreement with real PET-based diagnoses (accuracy = 0.80). Using synthetic PET, we developed a fully automatic AD diagnostic pipeline integrating PET synthesis and classification. The synthetic PET-based model (AUC = 0.78) outperforms T1-based (AUC = 0.68) and BBM-based (AUC = 0.73) models, while combining synthetic PET and BBMs further improved performance (AUC = 0.79). Ablation analysis supports the advantages of LLM integration and prompt engineering. Interpretation: Our language-enhanced generative model synthesizes realistic PET images, enhancing the utility of MRI and BBMs for Abeta spatial pattern assessment and improving the diagnostic workflow for Alzheimer's disease.

cs.CV↗

arXiv2025

EmbryoDiff: A Conditional Diffusion Framework with Multi-Focal Feature Fusion for Fine-Grained Embryo Developmental Stage Recognition

Identification of fine-grained embryo developmental stages during In Vitro Fertilization (IVF) is crucial for assessing embryo viability. Although recent deep learning methods have achieved promising accuracy, existing discriminative models fail to utilize the distributional prior of embryonic development to improve accuracy. Moreover, their reliance on single-focal information leads to incomplete embryonic representations, making them susceptible to feature ambiguity under cell occlusions. To address these limitations, we propose EmbryoDiff, a two-stage diffusion-based framework that formulates the task as a conditional sequence denoising process. Specifically, we first train and freeze a frame-level encoder to extract robust multi-focal features. In the second stage, we introduce a Multi-Focal Feature Fusion Strategy that aggregates information across focal planes to construct a 3D-aware morphological representation, effectively alleviating ambiguities arising from cell occlusions. Building on this fused representation, we derive complementary semantic and boundary cues and design a Hybrid Semantic-Boundary Condition Block to inject them into the diffusion-based denoising process, enabling accurate embryonic stage classification. Extensive experiments on two benchmark datasets show that our method achieves state-of-the-art results. Notably, with only a single denoising step, our model obtains the best average test performance, reaching 82.8% and 81.3% accuracy on the two datasets, respectively.

cs.CV↗

arXiv2025

MoGIC: Boosting Motion Generation via Intention Understanding and Visual Context

Existing text-driven motion generation methods often treat synthesis as a bidirectional mapping between language and motion, but remain limited in capturing the causal logic of action execution and the human intentions that drive behavior. The absence of visual grounding further restricts precision and personalization, as language alone cannot specify fine-grained spatiotemporal details. We propose MoGIC, a unified framework that integrates intention modeling and visual priors into multimodal motion synthesis. By jointly optimizing multimodal-conditioned motion generation and intention prediction, MoGIC uncovers latent human goals, leverages visual priors to enhance generation, and exhibits versatile multimodal generative capability. We further introduce a mixture-of-attention mechanism with adaptive scope to enable effective local alignment between conditional tokens and motion subsequences. To support this paradigm, we curate Mo440H, a 440-hour benchmark from 21 high-quality motion datasets. Experiments show that after finetuning, MoGIC reduces FID by 38.6\% on HumanML3D and 34.6\% on Mo440H, surpasses LLM-based methods in motion captioning with a lightweight text head, and further enables intention prediction and vision-conditioned generation, advancing controllable motion synthesis and intention understanding. The code is available at https://github.com/JunyuShi02/MoGIC

cs.CV↗

arXiv2025

Spatial-Aware Self-Supervision for Medical 3D Imaging with Multi-Granularity Observable Tasks

The application of self-supervised techniques has become increasingly prevalent within medical visualization tasks, primarily due to its capacity to mitigate the data scarcity prevalent in the healthcare sector. The majority of current works are influenced by designs originating in the generic 2D visual domain, which lack the intuitive demonstration of the model's learning process regarding 3D spatial knowledge. Consequently, these methods often fall short in terms of medical interpretability. We propose a method consisting of three sub-tasks to capture the spatially relevant semantics in medical 3D imaging. Their design adheres to observable principles to ensure interpretability, and minimize the performance loss caused thereby as much as possible. By leveraging the enhanced semantic depth offered by the extra dimension in 3D imaging, this approach incorporates multi-granularity spatial relationship modeling to maintain training stability. Experimental findings suggest that our approach is capable of delivering performance that is on par with current methodologies, while facilitating an intuitive understanding of the self-supervised learning process.

cs.CV↗

arXiv2025

SPRMamba: Surgical Phase Recognition for Endoscopic Submucosal Dissection with Mamba

Endoscopic Submucosal Dissection (ESD) is a minimally invasive procedure initially developed for early gastric cancer treatment and has expanded to address diverse gastrointestinal lesions. While computer-assisted surgery (CAS) systems enhance ESD precision and safety, their efficacy hinges on accurate real-time surgical phase recognition, a task complicated by ESD's inherent complexity, including heterogeneous lesion characteristics and dynamic tissue interactions. Existing video-based phase recognition algorithms, constrained by inefficient temporal context modeling, exhibit limited performance in capturing fine-grained phase transitions and long-range dependencies. To overcome these limitations, we propose SPRMamba, a novel framework integrating a Mamba-based architecture with a Scaled Residual TranMamba (SRTM) block to synergize long-term temporal modeling and localized detail extraction. SPRMamba further introduces the Hierarchical Sampling Strategy to optimize computational efficiency, enabling real-time processing critical for clinical deployment. Evaluated on the ESD385 dataset and the cholecystectomy benchmark Cholec80, SPRMamba achieves state-of-the-art performance (87.64% accuracy on ESD385, +1.0% over prior methods), demonstrating robust generalizability across surgical workflows. This advancement bridges the gap between computational efficiency and temporal sensitivity, offering a transformative tool for intraoperative guidance and skill assessment in ESD surgery. The code is accessible at https://github.com/Zxnyyyyy/SPRMamba.

cs.CV↗

arXiv2025

Intuitive Axial Augmentation Using Polar-Sine-Based Piecewise Distortion for Medical Slice-Wise Segmentation

Most data-driven models for medical image analysis rely on universal augmentations to improve accuracy. Experimental evidence has confirmed their effectiveness, but the unclear mechanism underlying them poses a barrier to the widespread acceptance and trust in such methods within the medical community. We revisit and acknowledge the unique characteristics of medical images apart from traditional digital images, and consequently, proposed a medical-specific augmentation algorithm that is more elastic and aligns well with radiology scan procedure. The method performs piecewise affine with sinusoidal distorted ray according to radius on polar coordinates, thus simulating uncertain postures of human lying flat on the scanning table. Our method could generate human visceral distribution without affecting the fundamental relative position on axial plane. Two non-adaptive algorithms, namely Meta-based Scan Table Removal and Similarity-Guided Parameter Search, are introduced to bolster robustness of our augmentation method. In contrast to other methodologies, our method is highlighted for its intuitive design and ease of understanding for medical professionals, thereby enhancing its applicability in clinical scenarios. Experiments show our method improves accuracy with two modality across multiple famous segmentation frameworks without requiring more data samples. Our preview code is available in: https://github.com/MGAMZ/PSBPD.

cs.CV↗

arXiv2025

Interpretable Auto Window Setting for Deep-Learning-Based CT Analysis

Whether during the early days of popularization or in the present, the window setting in Computed Tomography (CT) has always been an indispensable part of the CT analysis process. Although research has investigated the capabilities of CT multi-window fusion in enhancing neural networks, there remains a paucity of domain-invariant, intuitively interpretable methodologies for Auto Window Setting. In this work, we propose an plug-and-play module originate from Tanh activation function, which is compatible with mainstream deep learning architectures. Starting from the physical principles of CT, we adhere to the principle of interpretability to ensure the module's reliability for medical implementations. The domain-invariant design facilitates observation of the preference decisions rendered by the adaptive mechanism from a clinically intuitive perspective. This enables the proposed method to be understood not only by experts in neural networks but also garners higher trust from clinicians. We confirm the effectiveness of the proposed method in multiple open-source datasets, yielding 10%~200% Dice improvements on hard segment targets.

eess.IV↗