surveillance system developers

Abstract: Image compression is increasingly used before machine vision systems, yet standard codecs are mainly optimized for human perception rather than downstream task accuracy. We propose SCOPE (Semantic COdec Pre-Editing), a conditional image pre-editor that modifies images before compression to improve the bitrate and detection accuracy trade-off. The model is conditioned on the codec family and compression severity, allowing a single trained network to adapt to multiple compression settings without retraining or loading codec-specific weights. Instead of uniformly preserving the whole image, the pre-editor learns to protect task-relevant regions while simplifying less important background content to improve compressibility. The resulting rate-accuracy curves show consistent gains across different codec families and detector architectures, together with strong transferability to unseen learned and wavelet-based compression families.

Abstract: Point-supervised temporal action localization (PTAL) aims to recognize and localize action instances in untrimmed videos using only point-level annotations. Existing pseudo-label-based methods have made substantial progress in PTAL. However, confidence-based initialization can overlook well-localized proposals. Meanwhile, existing methods either keep pseudo-labels fixed during detector training or update them without fully exploiting alternative boundary evidence from the original proposal pool. These limitations constrain supervision quality, thereby hindering detection performance. To learn high-quality pseudo-labels, we propose a framework that integrates prior-guided initialization with consensus-guided refinement during detector training. Our framework comprises two modules: Prior-Guided Initialization (PGI) and Consensus-Guided Refinement (CGR). PGI combines proposal confidence with point-derived priors to establish instance-specific reliability for initial pseudo-label selection and candidate retention. CGR continually uses the retained original candidates to construct boundary targets, combining their consensus with evolving detector predictions to refine pseudo-labels throughout training. Experiments on four benchmarks demonstrate that our method achieves competitive detection performance compared with existing methods. In particular, our method achieves 68.5\% average mAP over tIoU thresholds 0.1--0.7 on THUMOS14, surpassing PseR by 3.7 percentage points. Code is available at https://github.com/zhenyingfang/LHQP

Abstract: Road anomalies such as collisions, violations, and sudden hazards demand more from video understanding than detection: a practical system must verify what happened, describe the scene, and explain the cause. We find these competences pull in opposite directions for vision-language models: closed-form questions reward precise, evidence-grounded reasoning, while open-ended answers are scored by similarity to reference text, where matching the annotation style matters as much as reasoning correctly. We present Reason-or-Recite (RoR), a two-specialist framework for the Traffic Anomaly Reasoning track of the 2026 AI City Challenge. RoR (1) rewrites the official chain-of-thought annotations into a structured, temporally grounded schema and LoRA-tunes a 9B vision-language model on the enhanced traces; (2) fine-tunes a 4-bit-quantized 27B model on answers alone, absorbing the annotation style; and (3) selects between the two candidates per question with a training-free LLM judge. RoR ranks first on the in-domain leaderboard with mean score 0.6788.

Abstract: This paper presents a novel strategy for UAV detection through visible (RGB) and infrared (IR) fusion, in order to improve accuracy and precision in harsh environments and weather conditions. In this research, two opposite approaches have been studied and implemented: the first concerns image-level (early) fusion, which merges two different spectra into a single image to which one detector is applied. On the other hand, the second approach, called decision-level (late) fusion, focuses on blending together the output bounding boxes of two different detection pipelines, aiming to create a common result that can be applied to both spectra simultaneously. Both strategies have been extensively tested in a dedicated environment, which included alternating low and high brightness scenes, different backgrounds (sky and ground), distances (5, 25 and 60 m), altitudes and zoom levels (1X, 2X and 4X). These tests allowed for the assessment of actual detection capabilities and system reliability in real-world applications. Furthermore, the two fusion strategies were evaluated and compared over a set of aligned images in order to highlight the better approach of the two. The results indicate that early fusion achieves better raw performance, since it incorporates the thermal information from the beginning of the training process. However, late fusion presents an innovative and useful detection workflow that can be applied in cases of sensor or frame misalignment. Both positive and negative aspects have also been pointed out and discussed, in order to give a more comprehensive overview. Finally, the paper outlines implementation considerations and practical recommendations based on operational priorities. Overall, the findings contribute to the research in RGB-IR fusion modalities, with the goal of enhancing UAV detection capabilities.

Abstract: Fine-grained alignment between text and video object trajectories is a core challenge in multimodal understanding. Existing embedding models excel at global retrieval, yet their single global representation entangles multiple objects and their spatiotemporal dynamics, limiting direct object-level grounding. To address this issue, we propose VTrajEmbed, an embedding model that decomposes a video into object trajectories and aligns each trajectory embedding with text for object-level grounding through retrieval. Our main contributions: (1)Slow-Fast Trajectory Encoding, a dual-branch architecture that establishes trajectory representations by densely sampling low-resolution frames for high-frequency motion cues and sparsely selecting high-resolution frames for scene context, aggregated by a shared LLM backbone; (2)Mask-Guided Contrastive Learning, a training objective built upon these representations that couples an intra-group loss and a batch-level bidirectional loss under a relation mask, excluding unverified relations and preventing confirmed matches from becoming false negatives; (3)Text2Traj-Bench, a trajectory-level text-video benchmark with model-assisted construction and human-verified query--trajectory relations. Experimental results show that VTrajEmbed-5B improves text-to-trajectory and trajectory-to-text R@1 over Qwen3-VL-Embedding-8B by 15.48 and 18.05 percentage points, respectively, and achieves 49.48% average R@1 across five global retrieval benchmarks, compared with 42.86% for Qwen3-VL-Embedding-8B.

Abstract: Atmospheric turbulence creates complex spatial and temporal distortions that significantly reduce image quality. This work compares two deep learning methods for turbulence reduction: Generative Adversarial Networks (GANs) and Transformer-based models. Both methods are trained and tested on the SOTIS dataset, which includes multi-frame sequences under different turbulence conditions with ground truth references. Our results show an important finding: GAN-based models perform better in severely degraded, low-quality situations, while Transformer-based models achieve better results in moderate to high-quality settings, especially in correcting geometric distortions. We also enhance the Turbulence Mitigation Transformer (TMT) by introducing a refined training strategy, an augmented input design that includes a temporally averaged frame as prior structural information, and a modified loss function that incorporates the Structural Similarity Index (SSIM). These improvements lead to consistent performance gains across various turbulence conditions. Overall, this study provides an in-depth evaluation of deep learning methods for turbulence reduction and offers practical guidance for selecting models in different imaging environments.

Abstract: Streaming video understanding requires answering questions from the observed video, with evidence drawn from recent observations or earlier events. Historical context can improve recall but may interfere with judgments based on recent frames. Its usefulness depends on both when it is introduced and which clips are selected, while relevance alone does not guarantee sufficient evidence for an answer. We propose SPARE, a framework that combines adaptive historical retrieval with visual encoder sharing. A lightweight gate predicts whether recent observations provide sufficient evidence to answer the question. When additional evidence is predicted to be necessary, a multimodal embedding model retrieves relevant historical clips to supplement the current context. LoRA instruction tuning trains the understanding model to use recent observations and retrieved history jointly and abstain when the evidence remains insufficient. To reduce redundant visual parameters, the understanding and retrieval branches share a visual encoder while retaining separate language backbones. Four linear adapters align the shared features with the retrieval model through distillation at the visual token and final retrieval embedding levels. With visual-only inputs, SPARE achieves 77.78% on the evaluated Real-time and Backward task groups of OVO-Bench and 73.31% on StreamingBench. Compared with Qwen3-VL-8B-Instruct evaluated at 2 fps, SPARE improves accuracy by 20.68 percentage points on OVO-Bench and 14.15 percentage points on StreamingBench.

Abstract: Streaming Video Large Language Models (VLMs) enable continuous video understanding, yet existing benchmarks focus on low-dynamic scenarios. Under bounded context budgets, models must balance temporal history, spatial resolution, and temporal granularity; sparse sampling at 1--2 FPS misses fast events. We introduce FastBench to evaluate high-dynamic perception in real-world video streams. Its trajectory-grounded pipeline combines QA generation from high-FPS clips, filtering of questions answerable at 2 FPS, answer verification using SAM3 and CoTracker3 trajectories, and three rounds of human inspection. FastBench contains 306 QA pairs across eight domains, six capabilities, and forward, instant, and backward temporal scopes, with human-annotated evidence intervals. We also present ProactiveFrame, a training-free baseline that adjusts incoming frame rates through text tokens. A dual-tier sliding window retains recent high-FPS observations while downsampling older ones into sparse history. Experiments reveal substantial limitations: the strongest model, Gemini-3.5-Flash, scores only 50.7%. Denser sampling improves Qwen3-VL-8B from 32.9% at 2 FPS to 44.6% at 24 FPS, but gains saturate as history is compressed. ProactiveFrame outperforms sparse uniform sampling by 5.4 and 1.5 percentage points, yet remains well below oracle-guided focusing, showing that current VLMs struggle to determine from the stream alone when finer temporal perception is needed. FastBench provides a testbed for high-dynamic streaming video understanding. Code and data: https://github.com/Ashone3/FastBench.

Abstract: N-version object detection (OD) is an approach to diversifying detection results using multiple models or input frames and reducing detection errors by aggregating individual results. Diversity and consistency across multiple detection results are critical information for characterizing the reliability of possible configurations of N-version OD systems. However, existing performance metrics such as mAP and Accuracy fail to capture these factors, as they are defined solely on the final outcome after aggregation. To overcome this limitation, we propose two reliability metrics particularly defined for N-version OD, namely coverage of errors in OD (Cov_OD) and certainty of accurate prediction in OD (Cer_OD), which can be computed from individual detection results without relying on voting strategies. We empirically demonstrate the unique features of the proposed metrics through a case study of N-version OD for a vehicle in an autonomous-driving simulator. We show that the proposed metrics can be used for 1) selecting version combinations with complementary error characteristics, 2) choosing an effective voting strategy based on diversity and consistency profiles, and 3) guiding incremental construction of N-version OD systems via pairwise two-version analysis. These results highlight the importance of the metrics that can guide the design of reliable N-version OD applications.

Abstract: Active robot vision requires controlling the camera to reveal task-relevant information that is hidden from the current viewpoint. For example, determining what is inside a box may require raising the camera and looking down into it. For viewpoint-dependent question answering, the challenge is to select camera motions that expose the visual evidence needed to answer the question. Although vision-language models (VLMs) can interpret observed images, selecting such motions requires anticipating the usefulness of unseen views. We quantify this usefulness as answerability, a VLM's estimate that a view suffices to answer the question, and present Rendering-Free Lookahead (RFL), a viewpoint-selection policy that ranks candidate camera motions by predicted future answerability. RFL transfers visual lookahead from deployment to offline training. At training, a privileged teacher renders candidate future views in 3D Gaussian Splatting (3DGS) scenes and uses a frozen VLM to compute one- and two-step answerability targets. Through two-stage distillation, a student learns to predict these action values from the question, recent visual observations, and a candidate camera motion. At deployment, RFL uses these predicted values to select camera motions without rendering future views. On 377 E3VS-Bench test episodes in unseen environments, RFL improves the mean judge score by 43\% over a direct-action baseline using the same VLM. These results support learning camera-control policies from privileged visual lookahead for viewpoint-dependent question answering.

Abstract: Roadside traffic reasoning requires every free-form textual claim to be backed by visual evidence. Existing grounded multimodal large language models (MLLMs) frequently exhibit say-point mismatch, in which the textual answer contradicts the bounding boxes the model localizes. Evaluation metrics that score answers and boxes separately leave this failure unpenalized. We trace the mismatch to the conventional answer-then-ground factorization, which commits to a numerical claim before any object is enumerated. To measure it, we build RoadSceneVQA-G, a benchmark of 34.7K question-answer pairs in which every free-form answer is linked to the set of boxes that witnesses it, and we propose the Answer-Grounding Consistency (AGC) evaluation suite. To address it, we introduce Enumerate-then-Answer (EtA), which reverses the generation order so that answer-evidence agreement becomes a property of the output structure, and Enumeration-Consistent Policy Optimization (ECPO), a reinforcement learning stage that uses the union of multiple rollouts as a recall teacher without ground-truth boxes. EtA raises say-point consistency from 26.6\% to 93.7\% and grounding F1 from 52.2\% to 73.0\%, and ECPO further increases F1 to 75.6\% without per-box supervision. On gRefCOCO, the same framework outperforms the strongest compared method, indicating that it transfers beyond traffic scenes. The project is available at \url{https://github.com/GuanRunwei/RoadSceneVQA-G}.

Abstract: Efficient long-video understanding with vision-language models (VLMs) is often framed as selecting informative frames or visual tokens at a fixed native resolution. We show that per-frame resolution can instead be traded for denser temporal coverage, while front-end decoding latency depends on the size of the candidate pool rather than the final token budget. An empirical study across multiple VLMs and long-video benchmarks yields three findings: dense low-resolution sampling outperforms sparse native-resolution sampling at matched token budgets; resolution-sensitive tasks benefit from selected high-resolution frames; and front-end decoding dominates wall time for hour-long videos. Motivated by these findings, we introduce LoHi, a training-free, single-pass framework that combines a dense low-resolution video stream with sparse high-resolution image frames through the VLM's native video and image pathways. LoHi-Anchor selects high-resolution frames using codec I-frame metadata, while LoHi-SemDiv uses query relevance and visual diversity over CLIP features. Across three long-video benchmarks, LoHi improves average accuracy by 10.6 percentage points over the native-resolution baseline at a matched token budget and by 5.2 percentage points over the strongest prior efficiency method. It also reduces front-end decoding latency by up to 7x on hour-long videos. Project page: https://sixundong.com/projects/lohi

Abstract: Memory-attention-based Video Instance Segmentation (VIS) methods have demonstrated strong zero-shot tracking capability, yet their substantial memory requirements confine them to short video clips and their single-prompt inference design makes multi-category open-vocabulary tracking computationally prohibitive. This work introduces two contributions toward fully automated tracking annotation of arbitrary video. The Generalized Presence Token (GPT) reformulates SAM3's inference pipeline to process N text prompts simultaneously via virtual prompt batching, reducing image encoding cost from O(N) to O(1) with no modifications to any learned component. The Annotation and Segmentation Handler (ASH) extends any memory-attention VIS tracker to sequences of arbitrary length through overlapping temporal chunks with IoU-based inter-chunk identity matching, requiring no dataset-specific training. Instantiated on SAM3, the resulting pipeline -- SAM3-ASH -- achieves state-of-the-art HOTA on MOTS20 under fully zero-shot conditions and remains competitive with trained specialists across seven additional benchmarks, while peak GPU memory consumption stays below 25 GB, establishing a practical baseline for scalable, training-free automated video annotation.

Abstract: Multimodal image fusion (MMIF) aims to integrate complementary information from different modalities into a high-quality fused image and support downstream tasks. Recently, feature decomposition has become an important paradigm by separating source images into common and modality-specific unique features. However, existing methods lack clear supervision because ground-truth (GT) decomposition feature maps are unavailable. They usually combine multiple image-level metrics as losses, which are inherently incomplete and may conflict since each pixel couples attributes such as texture, edge, and contour. To address this, we propose a 1D signal-level self-supervised feature decomposition paradigm. Our core insight is to reformulate feature decomposition from unclear 2D image-level supervision into an integral-driven 1D signal-level optimization problem. This objective-level reformulation uses the 1D signal form to compute the integral constraint. The decomposer is optimized by the integral area between common and original signals, enabling more stable optimization with a clear optimization objective. Our model follows a two-stage SSL framework. Stage I designs dual pretext tasks for integral-driven decomposition at the signal level and structure-preserving reconstruction at the image level. Stage II fuses unique features and combines them with common features to reconstruct the fused image. Experiments on representative MMIF tasks show state-of-the-art (SOTA) performance. Code: github.com/Wangjiayu0512/SIDFusion.

Abstract: Object re-identification (ReID) faces a recurring challenge: different identities can share highly similar global appearances, while the cues that distinguish them are localized, heterogeneous, and visible only under particular viewpoints. This challenge arises in animal ReID through markings, contours, and scars, in person ReID through subtle clothing and accessory cues, and in vehicle ReID through localized appearance details. Although visual foundation models encode such information in dense tokens, a single holistic descriptor can obscure discriminative local signals. We propose FM-ReID, an end-to-end framework that formulates local representation learning as selective competitive token routing. Its Competitive Fine-grained Mining module uses multiple mining queries and a residual query to compete for dense DINOv3 tokens. Above-prior selection retains tokens preferentially allocated to each mining query, while the residual slot receives tokens excluded from the retrieval descriptors. The resulting multi-query descriptors are jointly trained with a holistic representation for retrieval, without fixed spatial partitions or equal-area constraints. FM-ReID achieves strong results on animal, person, and vehicle ReID benchmarks, supporting competitive token routing as an effective way to augment holistic foundation-model representations.

Abstract: Continuous video counting requires distinguishing new observations from new objects or completed events. We introduce StaMina (State Maintenance), which learns to maintain counting state through state-conditioned updates. Recurrent visual context supports recognition; learned transitions maintain visibility, persistent identities, and completed-event records. A differentiable recurrence trains event transitions over legal paths constrained by count endpoints; visibility and association objectives train the object branch. A multi-source pipeline organizes 39.8K spatial queries and complementary event annotations into counting trajectories. On SVCBench, we evaluate counting adaptation with partial video overlap and held-out groups of linked annotations. Under prefix replay (Full) and persistent streaming (Stream), 4B and 8B models reach 41.9/36.4 and 44.9/38.2 Gaussian Precision Accuracy, respectively. The 8B model gains 10.9/3.2 points over Counting-SFT on the same queries. Matched-graph comparisons isolate phase conditioning and trajectory supervision, assessing training objectives alongside hard decisions. Online video benchmarks and count-conditioned decisions assess online understanding and task eligibility. Project Page: https://PLACEHOLDER.github.io/StaMina/

Abstract: Task-driven image restoration aims to improve both image quality and downstream task performance. However, existing methods predominantly focus on single degradation type and struggle to handle the diverse degradations encountered in real-world scenarios. Different degradations impose distinct restoration demands, and insufficient restoration may leave residual degradations and artifacts that impair object boundaries and semantic cues, thereby compromising downstream task performance. To address these challenges, we propose TaskIR, a two-stage task-driven unified image restoration framework that integrates degradation-adaptive restoration with task feedback refinement. In Stage I, a Degradation Representation Module (DRM) extracts degradation representations, enabling a Degradation-Guided Transformer Block (DGTB) to dynamically modulate feature transformations for adaptive restoration. In Stage II, a Task-to-Restoration Feedback Generation module (TRFG) transforms heterogeneous task features into restoration feedback by modeling task-representation discrepancies associated with the current restoration. Subsequently, a Selective Task Feedback Refinement module (STFR) assesses feedback relevance and selectively refines intermediate restoration features to mitigate interference with well-restored content. Extensive experiments demonstrate that TaskIR achieves competitive restoration quality and downstream task performance across diverse degradations and tasks.

Abstract: Reliable online video question answering requires tracking state transitions while selectively abstaining when visual evidence is insufficient. Existing benchmarks focus on static recognition or long-range retrieval, rarely evaluating these coupled capabilities under evolving and incomplete evidence. We present STORM-Bench, comprising 5,736 questions across 630 compact, change-dense episodes spanning five egocentric domains (STORM-Real) and two controlled simulation subsets (STORM-Sim) at 1 FPS. Questions are stratified by a proxy for accumulated change intensity (Low, Medium, High) and query-time answerability (Known, Uncertain). To measure reliability, we introduce STORM-BR, a harmonic metric over joint answer-status correctness that exposes abstention failures masked by aggregate accuracy, alongside STORM-BR-ATTR for uncertainty attribution. Across 14 video LLMs, online accuracy peaks at 60.3\% (mean 51.7\%), whereas STORM-BR ranges from 5.7\% to 35.6\% (mean 18.8\%), driven by pervasive overconfidence on uncertain queries. STORM-Bench shows that task accuracy masks these gaps in epistemic reliability and state tracking. Benchmark and code are available at https://github.com/siruzhong/STORM-Bench.

Abstract: Reliable uncertainty estimation is essential for deploying object detectors in autonomous systems operating in uncertain environments. Evidential Deep Learning (EDL) provides a principled framework for uncertainty-aware classification by representing network outputs as evidence and interpreting predictions through subjective logic. However, existing evidential object detectors typically combine evidential classification with regression uncertainty models that do not share the same theoretical foundation. In this work, we propose an evidential version of YOLOv8 in which both classification and bounding-box regression are formulated within a common evidential framework. Our approach exploits YOLOv8's distribution-based bounding-box representation, allowing the evidential formulation to be applied not only to classification but also to localisation. As a result, both tasks produce belief, uncertainty, and probability estimates that can be interpreted within the Dempster--Shafer framework. Experiments on KITTI, MUSES, and nuScenes show that the resulting detector remains broadly competitive with standard YOLOv8 in terms of detection accuracy while providing a localisation uncertainty that effectively discriminates between correct and erroneous detections. Moreover, this uncertainty becomes increasingly discriminative under domain shift.

Abstract: With face-recognition models now embedded in everyday authentication and surveillance, recent works have pinpointed a critical weakness: these models remain acutely vulnerable to adversarial semantic edits. I.e., adversarially produced semantic alterations to the input, such as slight aging or pose changes, can induce misclassifications. Certain existing attacks are powerful, but they can be computationally costly, rendering them inadequate for developing defenses (e.g., through adversarial training). To fill the gap, we introduce BoundStyle, a potent semantic attack operating in StyleGAN's rich latent space to maximize misclassification rates. Notably, BoundStyle achieves high attack success rates while being ${\sim}{\times}9.5$ faster than existing state-of-the-art attacks, making it suitable for adversarial training. Building on BoundStyle, we develop StyleAT, an efficient adversarial training scheme that incorporates low-budget attack variants yet defends against stronger and unseen semantic attacks. We evaluate on two datasets unseen during training and seven models, and find that StyleAT boosts robust accuracy against state-of-the-art attacks and outperforms common defenses in various settings.

Abstract: Camouflaged object detection (COD) is an important engineering task in intelligent optical perception, but it remains challenging when targets closely resemble their surroundings. Polarization imaging provides complementary physical cues, whereas existing methods typically assume fixed multimodal input configurations and entangle intra-polarization coordination with interaction between red-green-blue (RGB) and polarization representations. We propose LGFN, a lightweight gated RGB-polarization fusion framework supporting separately optimized RGB-only and polarization-assisted configurations. A deterministic Modality Router selects the appropriate configuration according to polarization availability. In the multimodal configuration, an availability-conditioned Modality Gate calibrates the available polarization branches; the Gated Polarization Hub coordinates learned degree of linear polarization (DoLP) and angle of polarization (AoP) representations with explicit polarization cues; and RGB-Polarization Cross Fusion introduces the coordinated representation into the RGB hierarchy through controlled residual interaction. The multimodal configuration requires neither sample-dependent statistics nor handcrafted quality descriptors during inference. On the complete 230-image PCOD_1200 test set, the RGB-only configuration achieves a mean absolute error of 0.0090, a Dice score of 0.8806, and an intersection over union of 0.8144, obtaining the best results on all six metrics among the evaluated RGB-based methods. Under a common local reevaluation protocol, the multimodal configuration outperforms PolarNet and IPNet on all six metrics. Relative to IPNet, it reduces the parameter count, floating-point operations, and latency by 53.1%, 73.6%, and 63.0%, respectively.

Abstract: Video Temporal Grounding (VTG) localizes the video segment that matches a natural-language query. Many queries describe an action performed by a particular entity. Existing methods often encode the query as a whole or use general video-text interactions, without explicitly checking whether the action and entity occur together. They may therefore select a segment that contains both concepts but not the event described by the query. We propose Interaction Aligned Action-Entity Video Temporal Grounding (IAE-VTG), which models this rela?tionship at both the representation and training assignment levels. First, the Fine-grained Disentangled Interaction Module (FDIM) separates action and entity related query information and aligns it with complementary motion and appearance features. It then combines token-level interactions to build representations that capture the relationship between the action and entity. Second, Interaction-Sensitive Assignment (ISA) adds this interaction evidence to bipartite matching, so training targets are selected using both temporal overlap and semantic compatibility. This reduces supervision from temporally plausible but semantically incorrect proposals. Experiments on QVHighlights, Charades?STA, and TACoS show that IAE-VTG consistently improves strong baselines and achieves competitive or state-of-the-art performance on standard grounding metrics. Additional analyses show that the method is especially effective when similar actions or entities appear at multiple times and produces more reliable assignments for complex events.

Abstract: Multi-object tracking (MOT) is an essential computer vision task that simultaneously tracks multiple objects in video sequences, with various applications in surveillance, autonomous navigation, and human-computer interaction. The tracking-by-detection (TBD) paradigm, which combines object detection with temporal association, has emerged as a leading approach, driven by innovative algorithms. Despite recent progress, fair evaluation of TBD-based methods remains a challenge. Many studies introduce modules such as similarity metrics, data association strategies, or motion models, but they are often evaluated under inconsistent protocols, with different baseline trackers, hyperparameters, and datasets. Such inconsistencies obscure the genuine contribution of each module and hinder objective comparison. This survey systematically reviews TBD-based MOT techniques, including similarity measurements, data association, camera motion compensation, and interpolation strategies. Starting from a minimal baseline tracker, we fairly evaluate the contributions of each method across diverse datasets and accumulate well-balanced methods. Our findings establish a strong baseline tracker and provide a foundation for the principled design of robust and versatile MOT systems suitable for real-world deployment.

Abstract: Spatio-temporal context has become increasingly crucial for visual tracking. However, most existing approaches extract spatio-temporal cues via discrete sampling strategies, which inherently deviate from the continuity of spatio-temporal context, thereby deteriorating tracking performance. To address this challenge, we propose TLCTrack, a novel tracking framework that models token-level spatio-temporal context through continuously updated salient tokens, enabling more accurate target representation. Specifically, TLCTrack incorporates three components: Masked Unidirectional Attention (MUA), Spatial Salient Token Collection (SSTC), and Temporal Salient Token Bank (TSTB) modules. By explicitly integrating spatio-temporal context, MUA extracts discriminative targetaware spatial features in the search region. To avoid the negative impact of background on feature learning, SSTC progressively suppresses background interference, thereby enhancing target spatial representation. Finally, TSTB captures high-quality spatio-temporal information through continuous salient token updates. Extensive experiments on five benchmarks demonstrate that our method achieves superior performance over state-of-the-art trackers. Code and models are available at https://github.com/xiading123/TLCTrack.