Detection, segmentation, depth and recognition research, plus the vision backbones that keep displacing the last generation.
arXiv:2511.16949v2 Announce Type: replace-cross Abstract: Dense 3D semantic occupancy perception is critical for mobile robots operating in pedestrian-rich environments, yet it remains underexplored...
arXiv:2604.16729v2 Announce Type: replace-cross Abstract: State-of-the-art large language models (LLMs) show high performance in general visual question answering. However, a fundamental limitation r...
arXiv:2609.00814v1 Announce Type: new Abstract: Remote sensing visual models have continuously advanced various interpretation tasks. However, the research process behind model improvement still heav...
The paper introduces ExiL, a mask‑conditioned progressive learning framework for bone ultrasound segmentation that models annotation as a structured refinement trajectory. ExiL uses a synthetic expert‑like brush simulator and a lightweight U‑Net to learn from imperfect masks, and it can be updated in real time from expert refinements. In experiments on UltraBones100k and a prospective volunteer dataset, ExiL cut average annotation time from 60 to 20 seconds per frame and improved mean Dice by about 0.045, achieving 0.87 Dice and 2.7 px boundary error with 10–50 ms inference.
arXiv:2609.00995v1 Announce Type: new Abstract: Recovering 3D digital humans from a single 2D image is an ill-posed computer vision problem due to the loss of depth information. Probabilistic 3D huma...
arXiv:2601.13913v3 Announce Type: replace Abstract: We consider monocular 3D human pose estimation (HPE), where the goal is to predict 3D human skeletal joints from a single 2D image, typically via 2...
The paper argues that in structured image classification, the key question is what information a classifier should receive before making a prediction, rather than which algorithm performs best. It proposes a systematic framework for constructing landmark-derived representations—such as coordinate, distance, angle, and hybrid features—and evaluates them on static hand gesture recognition. Experiments show that hybrid representations, which combine complementary geometric components, outperform raw coordinate features and other single-type representations, highlighting the importance of thoughtful feature construction.
arXiv:2609.01554v1 Announce Type: cross Abstract: Automated lesion segmentation in whole-body PET/CT is complicated by the variety of physiological tracer uptake patterns and by the differing appeara...
LatentPress compresses conversational histories and long documents into continuous memory tokens that a frozen decoder can read directly, eliminating the need for text reconstruction at inference. The method achieves 4–16× compression with only a small adapter (0.1% of the decoder’s parameters) and outperforms text summaries and OCR-based compression on LongMemEval and LongBench-QA benchmarks. Writing and reading are significantly faster than traditional text summarization or OCR reconstruction, demonstrating a practical machine-facing context interface beyond text and vision.
arXiv:2507.22052v3 Announce Type: replace Abstract: We present Ov3R, a novel framework for open-vocabulary semantic 3D reconstruction from RGB video streams, designed to advance Spatial AI. The syste...
arXiv:2510.26614v2 Announce Type: replace Abstract: We propose tokenization of events and present a tokenizer, Spiking Patches, specifically designed for event cameras. Given a stream of asynchronous...
MIDR (Multimodal Indexing for Document Retrieval) is a training‑free framework that enriches document indexes by converting rendered pages into verified textual fields with a multimodal LLM, then indexing those fields with BM25F and optionally fusing with dense retrieval. By shifting multimodal reasoning to index time, MIDR enables text‑centric serving while retaining multimodal evidence, achieving a 23.0% relative gain over BM25 on ViDoRe V3 and outperforming ColQwen2.5 on several domains with significantly smaller index memory and lower query latency.
The paper introduces ReuseRL, a method that applies the Minimum Description Length principle to agentic reinforcement learning. By extracting a shared skill dictionary from successful trajectories and adding a segmentation cost to the RL objective, ReuseRL discourages idiosyncratic behaviors and promotes reusable abstract patterns. Experiments on ALFWorld, TextWorld-Cooking, and Countdown-Stepwise show that ReuseRL improves both in‑distribution and out‑of‑distribution success compared to vanilla GRPO and other baselines.
HiLRP introduces a unified attribution framework for Vision Transformers (ViTs) that addresses the challenges posed by diverse architectural designs. By decomposing ViT operations into four basic types—linear maps, bilinear mixing, normalization/gating, and reindexing—HiLRP applies conservation‑satisfying relevance rules, enabling reliable explanations across a wide range of backbones. The method outperforms 14 existing attribution techniques on 10 architectures, maintaining conservation and improving localization accuracy (0.97 Pointing) compared to competitors.
arXiv:2509.25050v2 Announce Type: replace Abstract: Reinforcement Learning (RL) has emerged as a central paradigm for advancing Large Language Models (LLMs), where both pre-training and RL post-train...
arXiv:2609.01172v1 Announce Type: new Abstract: Monocular depth estimation has long stood as a fundamental challenge in computer vision, enabling a wide range of applications including 3D reconstruct...
The paper proposes a method called Restrict, Don't Retrain that enhances zero-shot aerial segmentation by using inference-time guidance from a vision‑language model (VLM). It combines a frozen foundation model that labels every pixel with two VLM queries: one to select relevant classes and another to locate small objects missed by the base model. Experiments on four aerial datasets show consistent performance gains at each stage where the base model is competent.
The paper introduces SARTM, a framework that adapts the Segment Anything Model (SAM) for RGB‑thermal (RGB‑T) semantic segmentation. It fine‑tunes SAM with LoRA layers, incorporates language guidance, and employs a Cross‑Modal Knowledge Distillation module to bridge modality gaps. The approach also modifies the segmentation head and adds an auxiliary semantic head, achieving superior performance on MFNET, PST900, and FMB benchmarks.
The paper introduces a cross‑modal pseudo‑labeling pipeline for unsupervised domain adaptation in semantic segmentation, particularly for waste sorting. It combines SAM for class‑agnostic region proposals with EVA‑CLIP to assign semantic labels via region‑text similarity, applying confidence filtering to ensure reliable pseudo‑labels for self‑training. An optional BLIP‑based language‑grounded verification further refines ambiguous regions, and the method shows consistent improvements over source‑only baselines on synthetic‑to‑real driving and lab‑to‑factory waste sorting shifts.
ViTAMINS is a method that incorporates synthetic hard negatives into unsupervised vision transformer pretraining to enhance representation quality. The approach is evaluated on ImageNet and a range of downstream tasks—including transfer learning, image retrieval, copy detection, and image/video segmentation—showing significant performance gains. The synthetic negatives also lead to emergent properties, such as representations that encode explicit semantic information and act as strong classifiers, improving over baselines by up to 11.3%.