Skip to main content
Search

# Object Detection

10 related articles

Subscribe to updates
  1. arXiv

    No labels on the target dataset: SAM, CLIP and a vision-language model team up for 60.8 AP zero-shot on COCO

    Human annotation on the target dataset is replaced by an automatic pipeline: SAM produces segmentation proposals, CLIP pre-filters and matches labels, and a large vision-language model labels only cluster medoid images, after which an open-set detector is trained on those pseudo-labels and refined iteratively. Under zero-shot settings on COCO val2017 and LVIS minival, the YOLO-backbone variant reaches 60.8 AP and 47.4 AP, while the Swin-T variant reaches 57.2 AP and 46.2 AP, indicating the gain is not only from the detector backbone.

  2. arXiv

    Tiny-object detectors stop guessing query counts: BUTR sets the budget from encoder confidence, reaching 35.7 AP with 316 queries on AI-TOD-v2

    In dense tiny-object detection, dynamic query designs previously entangled how many queries are decoded with which queries are kept, obscuring the source of gains. BUTR reframes dynamic query as allocating decoder query budget by object density: it sums encoder proposal confidences into a density proxy and maps it linearly to a per-image query count, keeping standard Top-k initialization without learned modules or heuristic filtering. On AI-TOD-v2, BUTR-L reaches 35.7 AP with 316 queries on average, versus 35.2 AP and 600 queries for same-scale D-FINE-L.

  3. Pothole Detectors Score 0.619 on the Published Split and 0.467 Without Near-Duplicates: A Video-Disjoint Re-Split of BharatPotHole

    The published split of the public road-defect dataset BharatPotHole overlaps between training and test: 38.8% of test images have a near-duplicate in training and 99.4% come from videos also used for training. After a deduplicated, video-disjoint re-split, the same YOLO26n detector reaches AP50 0.619 on the published test set and 0.467 on its near-duplicate-free subset, a paired difference of 0.152 (95% CI 0.090–0.234). In a controlled contamination test, replacing unseen test images with copies of training frames raised a video-disjoint model's AP50 from 0.133 to 0.494, while copies of unseen frames left it near 0.15.

  4. Visible and infrared stop being fused in one shot: staged modality control reaches 88.6% mAP50 on M3FD

    Visible and infrared images follow different imaging mechanisms, so direct concatenation or one-shot attention fusion can mistake background texture and thermal interference for complementary evidence. FCI-DETR splits detection into three linked stages: preserving modality-specific structure before fusion, calibrating cross-modal responses during fusion, and injecting infrared context by scale after fusion. It reaches 88.6%, 84.3%, 86.7% and 81.5% mAP50 on M3FD, DroneVehicle, VEDAI and FLIR, and beats a same-backbone RGB+IR RT-DETR baseline by 6.0 points on M3FD under a fixed protocol.

  5. arXiv

    A test set missing night images picks the wrong model: rewriting day photos into night restores the selection a fully covered set would make

    When a test set underrepresents a semantic condition such as night, model selection can be misled; a generative pipeline uses vision-language models to find the missing semantics and diffusion models to turn day images into night images, so the augmented test set selects the same model as a fully covered one.

  6. arXiv

    BRAHMa detects galactic bars with YOLO11x oriented boxes, reaching 1.14 kpc mean absolute error on 3150 Hoyle galaxies, close to the 0.99 kpc disagreement between two human catalogues

    The authors present BRAHMa, which converts Galaxy Zoo 3D volunteer bar masks into rotated-rectangle labels at a threshold calibrated against the Hoyle catalogue, fine-tunes a YOLO11x oriented-box detector on per-galaxy contrast-enhanced DESI Legacy images, and reproduces Hoyle bar lengths for 3150 galaxies with a mean absolute error of 1.14 kpc and Pearson 0.91 after a linear correction, close to the 0.99 kpc at which the two human catalogues agree, with no mask, threshold or fit at inference.

  7. OpenBox aligns 2D vision-foundation instance cues with 3D point clouds to produce more accurate and cheaper 3D box annotations without self-training across three autonomous-driving datasets

    OpenBox introduces a two-stage automatic annotation pipeline: in the first stage it uses a 2D vision foundation model and cross-modal instance alignment to associate instance-level cues from 2D images with the corresponding 3D point clouds; in the second stage it categorizes instances by rigidity and motion state and generates adaptive bounding boxes with class-specific size statistics, thereby producing high-quality 3D bounding box annotations without self-training and improving accuracy and efficiency over baselines on the Waymo Open Dataset, the Lyft Level 5 Perception dataset, and the nuScenes dataset.

  8. arXiv

    TAPe+ML v3 reaches 84.7 mAP50 and 65.3 mAP50-95 on COCO detection with under 100K parameters, and lifts 30-image industrial ore-blockage accuracy from about 28% with YOLO 26s to about 65%

    The work presents TAPe+ML v3, which replaces raw pixel tensors with a structured TAPe (Theory of Active Perception) representation and combines background, pointer, and prototype-clustering submodels under a coordinator; with fewer than 100,000 parameters it reports 84.7 mAP50 and 65.3 mAP50-95 on COCO detection, 80.7 mask mAP50 and 58.4 mask mAP50-95 on COCO instance segmentation, 88.1% ImageNet-1k Top-1 and 89.9% ImageNet-Real, 92% versus 47% on Imagenette under identical training with TAPe versus raw pixels, and an industrial ore-blockage pilot where backbone adaptation outperforms head-only adaptation.

  9. Development of an AI-Based Smartphone Application for Rapid Tick Identification and Geospatial Mapping: A Pilot Study

    This pilot study collected ticks between May 2023 and December 2024 from patients presenting tick bites at 10 medical institutions in Okayama, Hiroshima, and Kagawa prefectures, supplemented with wild tick images, and developed a two-stage AI pipeline of object detection and genus-level classification in which a YOLO-based model trained on 3258 annotated public images achieved mAP@0.5 of 0.954 and ResNet50 achieved mean validation accuracy of 95% on 533 tick images covering four genera, integrating the system into a prototype smartphone application with geospatial visualization to support clinical risk assessment and public health surveillance.

  10. LOCUS-Diff: Location-Controlled Thyroid Ultrasound Synthesis for Advancing Nodule Detection

    This work proposes LOCUS-Diff, a generative framework that combines a thyroid ultrasound synthesis foundation model, a nodule spatial control branch, and a COMB label-correction mechanism to produce synthetic samples judged by senior clinical experts in visual Turing tests as anatomically plausible and rivaling real scans, consistently outperforming the state of the art in downstream nodule detection on TN5000 and TN3k, and achieving higher mAP when augmenting a training subset with only 60% of real data with synthetic samples than training on the full real dataset.

You’re all caught up