论文检索

Jer Pelhan, Alan Lukežič, Matej Kristan

Few-shot detection-based counters estimate the number of instances in the image specified only by a few test-time exemplars. A common approach to localize objects across multiple sizes is to merge backbone features of different resolutions. Furthermore, to enable small object detection in densely populated regions, the input image is commonly upsampled and tiling is applied to cope with the increased computational and memory requirements. Because of these ad-hoc solutions, existing counters struggle with images containing diverse-sized objects and densely populated regions of small objects. We propose GeCo2, an end-to-end few-shot counting and detection method that explicitly addresses the object scale issues. A new dense query representation gradually aggregates exemplar-specific feature information across scales that leads to high-resolution dense queries that enable detection of large as well as small objects. GeCo2 surpasses state-of-the-art few-shot counters in counting as well as detection accuracy by ~10% while running ~3x faster at smaller GPU memory footprint.

Yucong Meng, Zhiwei Yang, Kexue Fu, Zhijian Song, Yonghong Shi

The challenge of accelerated MRI reconstruction lies in recovering high-quality images from undersampled k-space. Recently, the selective state space model (Mamba) has shown promising results in various tasks with balanced global receptive field and computational efficiency, shedding new light on MRI reconstruction. However, existing approaches directly flatten 2D images based on spatial positions and apply Mamba to vision tasks, failing to preserve and explore the content properties. In this paper, we posit that the key to unlocking Mamba's full potential for MRI reconstruction lies in content-aware sequence modeling. We investigate two fundamental challenges: (1) how to reasonably preserve semantic information when converting 2D images into 1D sequences, and (2) how to effectively identify and recover the crucial high-frequency textures. To this end, we introduce CAM, a novel framework that shifts Mamba-based MRI reconstruction from position-based to content-aware sequence modeling. Specifically, we introduce three modules: (1) the Semantic Preservation Scanning Module (SPSM) introduces learnable clustering centers to group similar pixels, establishing the semantic preserved sequence. (2) The Texture Extraction Scanning Module (TESM) acts as a differentiable local texture descriptor to estimate crucial high-frequency information, forming the texture emphasized sequence. (3) The Texture Enhancement Mamba Module (TEMM) further modulates the semantic sequence with texture-informed system matrices derived from the texture sequence, yielding both context- and texture-aware sequential representations. With these enhancements, CAM significantly outperforms existing methods across various datasets and under-sampling masks.

Yuxin Mao, Zhen Qin, Jinxing Zhou, Bin Fan, Jing Zhang, Yiran Zhong, Yuchao Dai

Vision Transformers (ViTs) have revolutionized computer vision, yet their self-attention mechanism lacks explicit spatial inductive biases, leading to suboptimal performance on spatially-structured tasks. Existing approaches introduce data-independent spatial decay based on fixed distance metrics, applying uniform attention weighting regardless of image content and limiting adaptability to diverse visual scenarios. Inspired by recent advances in large language models where content-aware gating mechanisms (e.g., GLA, HGRN2, FOX) significantly outperform static alternatives, we present the first successful adaptation of data-dependent spatial decay to 2D vision transformers. We introduce Spatial Decay Transformer (SDT), featuring a novel Context-Aware Gating (CAG) mechanism that generates dynamic, data-dependent decay for patch interactions. Our approach learns to modulate spatial attention based on both content relevance and spatial proximity. We address the fundamental challenge of 1D-to-2D adaptation through a unified spatial-content fusion framework that integrates manhattan distance-based spatial priors with learned content representations. Extensive experiments on ImageNet-1K classification and generation tasks demonstrate consistent improvements over strong baselines. Our work establishes data-dependent spatial decay as a new paradigm for enhancing spatial attention in vision transformers.

Liang Lv, Di Wang, Jing Zhang, Lefei Zhang

Semi-supervised semantic segmentation (S4) has advanced remote sensing (RS) analysis by leveraging unlabeled data through pseudo-labeling and consistency learning. However, existing S4 studies often rely on small-scale datasets and models, limiting their practical applicability. To address this, we propose S5, the first scalable framework for semi-supervised semantic segmentation in RS, which unlocks the potential of vast unlabeled Earth observation data typically underutilized due to costly pixel-level annotations. Built upon existing large-scale RS datasets, S5 introduces a data selection strategy that integrates entropy-based filtering and diversity expansion, resulting in the RS4P-1M dataset. Using this dataset, we systematically scale up S4 into a new pretraining paradigm, S4 pre-training (S4P), to pretrain RS foundation models (RSFMs) of varying sizes on this extensive corpus, significantly boosting their performance on land cover segmentation and object detection tasks. Furthermore, during fine-tuning, we incorporate a Mixture-of-Experts (MoE)-based multi-dataset fine-tuning approach, which enables efficient adaptation to multiple RS benchmarks with fewer parameters. This approach improves the generalization and versatility of RSFMs across diverse RS benchmarks. The resulting RSFMs achieve state-of-the-art performance across all benchmarks, underscoring the viability of scaling semi-supervised learning for RS applications.

Xiangyang Luo, Qingyu Li, Xiaokun Liu, Wenyu Qin, Miao Yang, Meng Wang, Pengfei Wan, Di Zhang, Kun Gai, Shao-Lun Huang

Current video generation models perform well at single-shot synthesis but struggle with multi-shot videos, facing critical challenges in maintaining character and background consistency across shots and flexibly generating videos of arbitrary length and shot count. To address these limitations, we introduce \textbf{FilmWeaver}, a novel framework designed to generate consistent, multi-shot videos of arbitrary length. First, it employs an autoregressive diffusion paradigm to achieve arbitrary-length video generation. To address the challenge of consistency, our key insight is to decouple the problem into inter-shot consistency and intra-shot coherence. We achieve this through a dual-level cache mechanism: a shot memory caches keyframes from preceding shots to maintain character and scene identity, while a temporal memory retains a history of frames from the current shot to ensure smooth, continuous motion. The proposed framework allows for flexible, multi-round user interaction to create multi-shot videos. Furthermore, due to this decoupled design, our method demonstrates high versatility by supporting downstream tasks such as multi-concept injection and video extension. To facilitate the training of our consistency-aware method, we also developed a comprehensive pipeline to construct a high-quality multi-shot video dataset. Extensive experimental results demonstrate that our method surpasses existing approaches on metrics for both consistency and aesthetic quality, opening up new possibilities for creating more consistent, controllable, and narrative-driven video content.

Tianyu Luan, Xuelu Feng, Zixin Zhu, Phani Nuney, Sheng Liu, Xuan Gong, David Doermann, Chunming Qiao, Junsong Yuan

Textured high-fidelity 3D models are crucial for games, AR/VR, and film, but human-aligned evaluation methods still fall behind despite recent advances in 3D reconstruction and generation. Existing metrics, such as Chamfer Distance, often fail to align with how humans evaluate the fidelity of 3D shapes. Recent learning-based metrics attempt to improve this by relying on rendered images and 2D image quality metrics. However, these approaches face limitations due to incomplete structural coverage and sensitivity to viewpoint choices. Moreover, most methods are trained on synthetic distortions, which differ significantly from real-world distortions, resulting in a domain gap. To address these challenges, we propose a new fidelity evaluation method that is based directly on 3D meshes with texture, without relying on rendering. Our method, named Textured Geometry Evaluation TGE, jointly uses the geometry and color information to calculate the fidelity of the input textured mesh with comparison to a reference colored shape. To train and evaluate our metric, we design a human-annotated dataset with real-world distortions. Experiments show that TGE outperforms rendering-based and geometry-only methods on real-world distortion dataset.

Zimao Lu, Hui Xu, Bing Liu, Ke Wang

Text-only training provides an attractive approach to address data scarcity challenges in zero-shot image captioning (ZIC), avoiding the expense of collecting paired image-text annotations. However, although these approaches perform well within training domains, they suffer from poor cross-domain generalization, often producing hallucinated content when encountering novel visual environments. Retrieval-based methods attempt to mitigate this limitation by leveraging external knowledge, but they can paradoxically exacerbate hallucination when retrieved captions contain entities irrelevant to the inputs. We introduce the concept of negative entities—objects that appear in generated caption but are absent from the input—and propose Negative Entity Suppression (NES) to tackle this challenge. NES seamlessly integrates three stages: (1) it employs synthetic images to ensure consistent image-to-text retrieval across both training and inference; (2) it filters negative entities from retrieved content to enhance accuracy; and (3) it applies attention-level suppression using identified negative entities to further minimize the impact of hallucination-prone features. Evaluation across multiple benchmarks demonstrates that NES maintains competitive in-domain performance while improving cross-domain transfer and reducing hallucination rates, achieving new state-of-the-art results in ZIC.

Yuanyuan Liu, Haiyang Mei, Dongyang Zhan, Jiayue Zhao, Dongsheng Zhou, Bo Dong, Xin Yang

3D visual grounding (3DVG) identifies objects in 3D scenes from language descriptions. Existing zero-shot approaches leverage 2D vision–language models (VLMs) by converting 3D spatial information (SI) into forms amenable to VLM processing, typically as composite inputs such as specified-view renderings or video sequences with overlaid object markers. However, this VLM ⊕ SI paradigm yields entangled visual representations that compel the VLM to process entire cluttered cues, making it hard to exploit spatial–semantic relationships effectively. In this work, we propose a new VLM ⊗ SI paradigm that externalizes the 3D SI into a form enabling the VLM to incrementally retrieve only what it needs during reasoning. We instantiate this paradigm with a novel View-on-Graph (VoG) method, which organizes the scene into a multi-modal, multi-layer scene graph and allows the VLM to operate as an active agent that selectively accesses necessary cues as it traverses the scene. This design offers two intrinsic advantages: (i) by structuring 3D context into a spatially and semantically coherent scene graph rather than confounding the VLM with densely entangled visual inputs, it lowers the VLM's reasoning difficulty; and (ii) by actively exploring and reasoning over the scene graph, it naturally produces transparent, step-by-step traces for interpretable 3DVG. Extensive experiments show that VoG achieves state-of-the-art zero-shot performance, establishing structured scene exploration as a promising strategy for advancing zero-shot 3DVG.

Chen Liu, Wen Li, Yongshu Huang, Minghang Zhu, Yuyang Yang, Dunqiang Liu, Sheng Ao, Cheng Wang

LiDAR odometry is a critical component of SLAM in autonomous driving and robotics. Learning-based methods have shown remarkable performance by regressing relative poses in an end-to-end manner. However, when applying these trained models, originally developed on the widely used KITTI dataset, to other scenes, performance often drops significantly. In other words, existing methods struggle to generalize well to new environments. To address this challenge, we propose RCP-LO, a simple yet effective LiDAR odometry framework. We introduce a novel representation for relative poses, reformulating them as relative coordinates, which can then be solved using geometrical verification. This approach avoids overly simplified pose representations and makes better use of scene geometry, thereby improving generalization. Moreover, to capture the inherent uncertainties in relative pose estimation from occluded LiDAR point clouds from dynamic environments, we adapt our framework to learn a denoising diffusion model, allowing for sampling plausible relative coordinates while improving robustness. We also introduce a differentiable geometric weighted singular value decomposition module, enabling efficient pose estimation through a single forward pass. Extensive experiments demonstrate that RCP-LO, trained exclusively on the KITTI dataset, achieves competitive performance compared to SOTA learning-based methods and generalizes effectively to the KITTI-360, Ford, and Oxford datasets.

Wenkai Lin, Qiming Xia, Wen Li, Xun Huang, Chenglu Wen

Multi-agents rely on accurate poses to share and align observations, enabling a collaborative perception of the environment. However, traditional GNSS-based localization often fails in GNSS-denied environments, making consistent feature alignment difficult in collaboration. To tackle this challenge, we propose a robust GNSS-free collaborative perception framework based on LiDAR localization. Specifically, we propose a lightweight Pose Generator with Confidence (PGC) to estimate compact pose and confidence representations. To alleviate the effects of localization errors, we further develop the Pose-Aware Spatio-Temporal Alignment Transformer (PASTAT), which performs confidence-aware spatial alignment while capturing essential temporal context. Additionally, we present a new simulation dataset, V2VLoc, which can be adapted for both LiDAR localization and collaborative detection tasks. V2VLoc comprises three subsets: Town1Loc, Town4Loc, and V2VDet. Town1Loc and Town4Loc offer multi-traversal sequences for training in localization tasks, whereas V2VDet is specifically intended for the collaborative detection task. Extensive experiments conducted on the V2VLoc dataset demonstrate that our approach achieves state-of-the-art performance under GNSS-denied conditions. We further conduct extended experiments on the real-world V2V4Real dataset to validate the effectiveness and generalizability of PASTAT.

Rifen Lin, Alex Jinpeng Wang, Jiawei Mo, Min Li

Multimodal pretraining has revolutionized visual understanding, but its impact on video-based person re-identification (ReID) remains underexplored. Existing approaches often rely on video-text pairs, yet suffer from two fundamental limitations: (1) lack of genuine multimodal pretraining, and (2) text poorly captures fine-grained temporal motion—an essential cue for distinguishing identities in video. In this work, we take a bold departure from text-based paradigms by introducing the first skeleton-driven pretraining framework for ReID. To achieve this, we propose Contrastive Skeleton-Image Pretraining for ReID (CSIP-ReID), a novel two-stage method that leverages skeleton sequences as a spatiotemporally informative modality aligned with video frames. In the first stage, we employ contrastive learning to align skeleton and visual features at sequence level. In the second stage, we introduce a dynamic Prototype Fusion Updater (PFU) to refine multimodal identity prototypes, fusing motion and appearance cues. Moreover, we propose a Skeleton Guided Temporal Modeling (SGTM) module that distills temporal cues from skeleton data and integrates them into visual features. Extensive experiments demonstrate that CSIP-ReID achieves new state-of-the-art results on standard video ReID benchmarks (MARS, LS-VID, iLIDS-VID). Moreover, it exhibits strong generalization to skeleton-only ReID tasks (BIWI, IAS), significantly outperforming previous methods. CSIP-ReID pioneers an annotation-free and motion-aware pretraining paradigm for ReID, opening a new frontier in multimodal representation learning.

Xiaoye Liang, Lai Jiang, Minglang Qiao, Yichen Guo, Yue Zhang, Xin Deng, Shengxi Li, Yufan Liu, Mai Xu

In recent years, the development of burst imaging technology has improved the capture and processing capabilities of visual data, enabling a wide range of applications. However, the redundancy in burst images leads to the increased storage and transmission demands, as well as reduced efficiency of downstream tasks. To address this, we propose a new task of Burst Image Quality Assessment (BuIQA), to evaluate the task-driven quality of each frame within a burst sequence, providing reasonable cues for burst image selection. Specifically, we establish the first benchmark dataset for BuIQA, consisting of 7,346 burst sequences with 45,827 images and 191,572 annotated quality scores for multiple downstream scenarios. Inspired by the data analysis, a unified BuIQA framework is proposed to achieve an efficient adaption for BuIQA under diverse downstream scenarios. Specifically, a task-driven prompt generation network is developed with heterogeneous knowledge distillation, to learn the priors of the downstream task. Then, the task-aware quality assessment network is introduced to assess the burst image quality based on the task prompt. Extensive experiments across 10 downstream scenarios demonstrate the impressive BuIQA performance of the proposed approach, outperforming the state-of-the-art. Furthermore, it can achieve 0.33 dB PSNR improvement in the downstream tasks of denoising and super-resolution, by applying our approach to select the high-quality burst frames.

Yuxuan Li, Xiang Li, Yunheng Li, Yicheng Zhang, Yimian Dai, Qibin Hou, Ming-Ming Cheng, Jian Yang

With the rapid advancement of remote sensing technology, high-resolution multi-modal imagery is now more widely accessible. Conventional object detection models are trained on a single dataset, often restricted to a specific imaging modality and annotation format. However, such an approach overlooks the valuable shared knowledge across multi-modalities and limits the model’s applicability in more versatile scenarios. This paper introduces a new task called Multi-Modal Datasets and Multi-Task Object Detection (M2Det) for remote sensing, designed to accurately detect horizontal or oriented objects from any sensor modality. This task poses challenges due to 1) the trade-offs involved in managing multi-modal modelling and 2) the complexities of multi-task optimization. To address these, we establish a benchmark dataset and propose a unified model, SM3Det (Single Model for Multi-Modal datasets and Multi-Task object Detection). SM3Det leverages a grid-level sparse MoE backbone to enable joint knowledge learning while preserving distinct feature representations for different modalities. Furthermore, we propose a novel consistency and synchronization optimization mechanism, allowing it to effectively handle varying levels of learning difficulty across modalities and tasks. Extensive experiments demonstrate SM3Det's effectiveness and generalizability, consistently outperforming the combination of specialized models on individual datasets.

Yifan Li, Kun Zhou, Xin Zhao, Lei Fang, Jirong Wen

As scaling up training data has significantly improved the general multimodal capabilities of Large Vision-Language Models (LVLMs), they still suffer from the hallucination issue, generating text that is inconsistent with the visual input. This phenomenon motivates us to systematically investigate the role of training data in hallucination. We introduce a new benchmark, POPEv2, which consists of counterfactual images collected from the training data of LVLMs with certain objects masked. Through comprehensive evaluation on POPEv2, we find that current LVLMs suffer from training bias: they fail to fully leverage their training data and hallucinate more frequently on images seen during training. Specifically, they perform poorly on counterfactual images, often incorrectly answering “Yes” to questions about masked objects. To understand this issue, we conduct probing experiments on the models’ internal components, revealing that this training bias is primarily located in the language modeling (LM) head, which fails to correctly translate accurate visual representations into textual outputs. Based on these findings, we propose Obliviate, an efficient and lightweight unlearning method designed to mitigate object hallucination via training bias unlearning. Obliviate identifies the discrepancy between ground-truth labels and model outputs on the training data as a proxy for bias and adopts a parameter- and data-efficient fine-tuning strategy that only updates the LM head. Extensive experiments demonstrate the effectiveness of our approach. While only reusing the training data and updating approximately 2% of the parameters, Obliviate significantly reduces hallucination across both discriminative and generative tasks. Furthermore, it demonstrates strong scalability with respect to both model size (2B to 72B) and training data volume, and exhibits promising generalization to hallucination types beyond object-level hallucination.

Yanshu Li, Jianjiang Yang, Zhennan Shen, Ligong Han, Haoyan Xu, Ruixiang Tang

Modern large vision-language models (LVLMs) convert each input image into a large set of tokens that far outnumber the text tokens. Although this improves visual perception, it also introduces severe image token redundancy. Because image tokens contain sparse information, many contribute little to reasoning but greatly increase inference cost. Recent image token pruning methods address this issue by identifying important tokens and removing the rest. These methods improve efficiency with only small performance drops. However, most of them focus on single-image tasks and overlook multimodal in-context learning (ICL), where redundancy is higher and efficiency is more important. Redundant tokens weaken the advantage of multimodal ICL for rapid domain adaptation and lead to unstable performance. When existing pruning methods are applied in this setting, they cause large accuracy drops, which exposes a clear gap and the need for new approaches. To address this, we propose Contextually Adaptive Token Pruning (CATP), a training-free pruning method designed for multimodal ICL. CATP uses two stages of progressive pruning that fully reflect the complex cross-modal interactions in the input sequence. After removing 77.8% of the image tokens, CATP achieves an average performance gain of 0.6% over the vanilla model on four LVLMs and eight benchmarks, clearly outperforming all baselines. At the same time, it improves efficiency by reducing inference latency by an average of 10.78%. CATP strengthens the practical value of multimodal ICL and lays the foundation for future progress in interleaved image-text settings.

Yangwei Li, Xiaochuan Wang, Xin Shang, Haisheng Li

Point cloud quality assessment (PCQA) has advanced significantly with synthetic datasets offering diverse distortion coverage for model training. However, when applied to new application scenarios, models often suffer from performance drops due to mismatched distortion characteristics between source and target domains. Most current methods use all available synthetic distortions, which may introduce irrelevant features and hinder generalization. To address this, we propose DST-PCQA, a distortion-selective training framework for PCQA. Unlike previous approaches that treat all distortions equally, DST-PCQA identifies and selects distortion types most relevant to a target domain by analyzing inter-domain distortion similarity. This selective strategy reduces negative transfer and enables efficient domain-specific training. To fully leverage the selected distortions for both classification and quality prediction, we adopt a dual-branch architecture that fuses 2D visual cues and 3D geometric structure via cross-modal attention. This design supports multi-level feature alignment across modalities and enables fine-grained distortion understanding. Extensive evaluations across three target domains have verified the effectiveness of DST-PCQA over full-set training baselines. Moreover, its distortion-selective strategy is orthogonal to existing model-based PCQA methods, enabling improved cross-domain performance and reduced training costs across a wide range of architectures.

Shuangzhi Li, Junlong Shen, Lei Ma, Xingyu Li

LiDAR-based 3D object detection models often struggle to generalize to real-world environments due to limited object diversity in existing datasets. To tackle it, we introduce the first generalized cross-domain few-shot (GCFS) task in 3D object detection, aiming to adapt a source-pretrained model to both common and novel classes in a new domain with only few-shot annotations. We propose a unified framework that learns stable target semantics under limited supervision by bridging 2D open-set semantics with 3D spatial reasoning. Specifically, an image-guided multi-modal fusion injects transferable 2D semantic cues into the 3D pipeline via vision-language models, while a physically-aware box search enhances 2D-to-3D alignment via LiDAR priors. To capture class-specific semantics from sparse data, we further introduce contrastive-enhanced prototype learning, which encodes few-shot instances into discriminative semantic anchors and stabilizes representation learning. Extensive experiments on GCFS benchmarks demonstrate the effectiveness and generality of our approach in realistic deployment settings.

Jiahao Li, Yusheng Luo, Yunzhong Lou, Xiangdong Zhou

We present ReCAD, a reinforcement learning (RL) framework that bootstraps pretrained large models (PLMs) to generate precise parametric computer-aided design (CAD) models from multimodal inputs by leveraging their inherent generative capabilities. With just access to simple functional interfaces (e.g., point coordinates), our approach enables the emergence of complex CAD operations (e.g., pattern replication and mirror). This stands in contrast to previous methods, which typically rely on knowledge injected through supervised fine-tuning (SFT), offer limited support for editability, and fail to exploit the strong generative priors of PLMs. Specifically, the ReCAD framework begins by fine-tuning vision-language models (VLMs) to equip them with basic CAD model generation capabilities, where we rewrite CAD scripts into parameterized code that is leveraged to generate accurate textual descriptions for supervision. Then, we propose a novel RL strategy that incorporates parameterized code as guidance to enhance the model’s reasoning on challenging questions. Furthermore, we employ a hierarchical primitive learning process to progressively teach structured and compositional skills under a unified reward function that ensures both geometric accuracy and semantic fidelity. ReCAD sets a new state-of-the-art in both text-to-CAD and image-to-CAD tasks, significantly improving geometric accuracy across in-distribution and out-of-distribution settings. In the image-to-CAD task, for instance, it reduces the mean Chamfer Distance from 73.47 to 29.61 (in-distribution) and from 272.06 to 80.23 (out-of-distribution), outperforming existing baselines by a substantial margin.

Chunming Li, Shidong Wang, Haofeng Zhang

To address the limitations of transductive learning in evolving real-world scenarios where unknown categories may continuously emerge, Continual Generalized Category Discovery (C-GCD) presents a novel paradigm that extends conventional category discovery frameworks. Unlike traditional static learning environments, C-GCD requires models to incrementally discover novel categories across multiple operational phases while maintaining discrimination capabilities for previously learned classes, posing significant challenges in balancing stability and plasticity. Prior approaches typically employ parameter-level knowledge distillation from historical models to alleviate catastrophic forgetting, which effectively preserves prior knowledge and optimizes computational efficiency. However, our analysis reveals that the persistent availability of samples from previous stages enables more sophisticated knowledge preservation strategies. Specifically, we present a Fix and Explore strategy that employs distinct learning methodologies for different types of potential data, aiming to preserve the features of old categories as much as possible and gradually exploring the potential distribution of new class latent spaces, we can enhance the model's ability to discover novel categories. This paper investigates this effect and introduces a novel heuristic paradigm to solve the C-GCD problem, called Fix and Explore (FaE), which aims to provide sufficient imaginative space for new classes while preserving the classification ability for old tasks. We conducted experiments across multiple datasets and performed detailed comparisons. The results demonstrate that our method achieves state-of-the-art performance at each stage across all datasets.

Hojun Lee, Mijin Koo, Yeji Song, Nojun Kwak

Recent advancements in diffusion models have made fine-tuning text-to-image models for personalization increasingly accessible, but have also raised significant concerns regarding unauthorized data usage and privacy infringement. Current protection methods are limited to passively degrading image quality, failing to achieve stable control. While Targeted Data Protection (TDP) offers a promising paradigm for active redirection toward user-specified target concepts, existing TDP attempts suffer from poor controllability due to snapshot-matching approaches that fail to account for complete learning dynamics. We introduce TAFAP (Trajectory Alignment via Fine-tuning with Adversarial Perturbations), the first method to successfully achieve effective TDP by controlling the entire training trajectory. Unlike snapshot-based methods whose protective influence is easily diluted as training progresses, TAFAP employs trajectory-matching inspired by dataset distillation to enforce persistent, verifiable transformations throughout fine-tuning. We validate our method through extensive experiments, demonstrating the first successful targeted transformation in diffusion models with simultaneous control over both identity and visual patterns. TAFAP significantly outperforms existing TDP attempts, achieving robust redirection toward target concepts while maintaining high image quality. This work enables verifiable safeguards and provides a new framework for controlling and tracing alterations in diffusion model outputs.