Frozen Vision Transformers for Dense Prediction on Small Datasets: A Case Study in Arrow Localization

• URL: http://arxiv.org/abs/2604.16758v1
• Date: Sat, 18 Apr 2026 00:17:48 GMT
• Title: Frozen Vision Transformers for Dense Prediction on Small Datasets: A Case Study in Arrow Localization
• Abstract summary: We present a system for automated detection, localization, and scoring of arrow punctures on 40,cm indoor archery target faces.<n>Our pipeline combines three components: a color-based canonical rectification stage, a frozen self-supervised vision transformer, and guided feature upsampling.<n>Across three cross-validation folds, we achieve a mean F1 score of $0.893 pm 0.011$ and a mean localization error of $1.41 pm 0.06$,mm.
• Score: 0.0
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: We present a system for automated detection, localization, and scoring of arrow punctures on 40\,cm indoor archery target faces, trained on only 48 annotated photographs (5{,}084 punctures). Our pipeline combines three components: a color-based canonical rectification stage that maps perspective-distorted photographs into a standardized coordinate system where pixel distances correspond to known physical measurements; a frozen self-supervised vision transformer (DINOv3 ViT-L/16) paired with AnyUp guided feature upsampling to recover sub-millimeter spatial precision from $32 \times 32$ patch tokens; and lightweight CenterNet-style detection heads for arrow-center heatmap prediction. Only 3.8\,M of 308\,M total parameters are trainable. Across three cross-validation folds, we achieve a mean F1 score of $0.893 \pm 0.011$ and a mean localization error of $1.41 \pm 0.06$\,mm, comparable to or better than prior fully-supervised approaches that require substantially more training data. An ablation study shows that the CenterNet offset regression head, typically essential for sub-pixel refinement, provides negligible detection improvement while degrading localization in our setting. This suggests that guided feature upsampling already resolves the spatial precision lost through patch tokenization. On downstream archery metrics, the system recovers per-image average arrow scores with a median error of 1.8\% and group centroid positions to within a median of 4.00\,mm. These results demonstrate that frozen foundation models with minimal task-specific adaptation offer a practical paradigm for dense prediction in small-data regimes.

Related papers

• DepthART: Scaling Foundation Monocular Depth to Tiny Models [51.10174763031444]
  DepthART is a compact MDE model for on-device deployment across diverse scenes.<n>We show how DepthART consistently surpasses previous tiny baselines in both zero-shot generalization and metric accuracy.<n>We also provide a scalable model family, with DepthART-S reaching 347/245 FPS (strict FP32) on an GTX A6000 at $2242/4482$, 102 FPS (TF32) on a Orin NX 8GB, and over 15 FPS (FP32) on a Jetson Nano 4GB.
  arXiv  Detail & Related papers  (2026-07-19T06:54:52Z)
• FoundationGeo: Learning Spatial Pixel-Wise Fields for Monocular Metric Geometry [52.232826445242644]
  FoundationGeo bridges relative and metric prediction via spatial calibration and principled data design.<n>Stage 1 learns a high-fidelity, affine-invariant geometry model by initializing with DINOv3.<n>Stage 2 moves beyond global scaling by introducing pixel-wise calibration fields for metric estimation.
  arXiv  Detail & Related papers  (2026-07-13T14:10:01Z)
• Learning to Navigate Efficiently with Only 0.58M Trainable Parameters [12.68304107799412]
  We propose a navigation model in which operations with known closed-form structure are computed analytically and serve as interfaces between three small learned modules.<n>The system trains only 0.58M out of a total of 22.7M parameters, on 44k frames in under one GPU-hour, yet approaches the performance of state-of-the-art models on navigation tasks.
  arXiv  Detail & Related papers  (2026-07-13T02:49:55Z)
• Graph Theoretical Outlier Rejection for 4D Radar Registration in Feature-Poor Environments [0.6171990546748666]
  Automotive 4D imaging radar is well suited for operation in dusty and low-visibility environments.<n>But registration remains challenging due to scan sparsity and spurious detections caused by noise and multipath reflections.<n>We propose a radar-adapted pairwise distance-invariant scoring function for clique-based graph (PCM)<n>Our method reduces segment relative position error (RPE) by 29.6% on 1 m segments and by up to on 100 m segments.
  arXiv  Detail & Related papers  (2026-04-16T10:44:47Z)
• Zero-shot Inexact CAD Model Alignment from a Single Image [53.37898107159792]
  A practical approach to infer 3D scene structure from a single image is to retrieve a closely matching 3D model from a database and align it with the object in the image.<n>Existing methods rely on supervised training with images and pose annotations, which limits them to a narrow set of object categories.<n>We propose a weakly supervised 9-DoF alignment method for inexact 3D models that requires no pose annotations and generalizes to unseen categories.
  arXiv  Detail & Related papers  (2025-07-04T04:46:59Z)
• FocusTune: Tuning Visual Localization through Focus-Guided Sampling [61.79440120153917]
  FocusTune is a focus-guided sampling technique to improve the performance of visual localization algorithms. We demonstrate that FocusTune both improves or matches state-of-the-art performance whilst keeping ACE's appealing low storage and compute requirements. This combination of high performance and low compute and storage requirements is particularly promising for applications in areas like mobile robotics and augmented reality.
  arXiv  Detail & Related papers  (2023-11-06T04:58:47Z)
• Fixation-based Self-calibration for Eye Tracking in VR Headsets [0.21561701531034413]
  The proposed method is based on the assumptions that the user's viewpoint can freely move. fixations are first detected from the time-series data of uncalibrated gaze directions. The calibration parameters are optimized by minimizing the sum of a dispersion metrics of the PoRs.
  arXiv  Detail & Related papers  (2023-11-01T09:34:15Z)
• Coordinate Transformer: Achieving Single-stage Multi-person Mesh Recovery from Videos [91.44553585470688]
  Multi-person 3D mesh recovery from videos is a critical first step towards automatic perception of group behavior in virtual reality, physical therapy and beyond. We propose the Coordinate transFormer (CoordFormer) that directly models multi-person spatial-temporal relations and simultaneously performs multi-mesh recovery in an end-to-end manner. Experiments on the 3DPW dataset demonstrate that CoordFormer significantly improves the state-of-the-art, outperforming the previously best results by 4.2%, 8.8% and 4.7% according to the MPJPE, PAMPJPE, and PVE metrics, respectively.
  arXiv  Detail & Related papers  (2023-08-20T18:23:07Z)
• Convolutional Cross-View Pose Estimation [9.599356978682108]
  We propose a novel end-to-end method for cross-view pose estimation. Our method is validated on the VIGOR and KITTI datasets. On the Oxford RobotCar dataset, our method can reliably estimate the ego-vehicle's pose over time.
  arXiv  Detail & Related papers  (2023-03-09T13:52:28Z)
• Centroid-centered Modeling for Efficient Vision Transformer Pre-training [44.24223088955106]
  Masked Image Modeling (MIM) is a new self-supervised vision pre-training paradigm using a Vision Transformer (ViT) Our proposed centroid-based approach, CCViT, leverages k-means clustering to obtain centroids for image modeling without supervised training of the tokenizer model. Our approach achieves competitive results with recent baselines without external supervision and distillation training from other models.
  arXiv  Detail & Related papers  (2023-03-08T15:34:57Z)
• Structure-Aware Long Short-Term Memory Network for 3D Cephalometric Landmark Detection [37.031819721889676]
  We propose a novel Structure-Aware Long Short-Term Memory framework (SA-LSTM) for efficient and accurate 3D landmark detection. SA-LSTM first locates the coarse landmarks via heatmap regression on a down-sampled CBCT volume. It then progressively refines landmarks by attentive offset regression using high-resolution cropped patches. Experiments show that our method significantly outperforms state-of-the-art methods in terms of efficiency and accuracy.
  arXiv  Detail & Related papers  (2021-07-21T06:35:52Z)
• Soft Expectation and Deep Maximization for Image Feature Detection [68.8204255655161]
  We propose SEDM, an iterative semi-supervised learning process that flips the question and first looks for repeatable 3D points, then trains a detector to localize them in image space. Our results show that this new model trained using SEDM is able to better localize the underlying 3D points in a scene.
  arXiv  Detail & Related papers  (2021-04-21T00:35:32Z)
• Generalized Focal Loss: Learning Qualified and Distributed Bounding Boxes for Dense Object Detection [85.53263670166304]
  One-stage detector basically formulates object detection as dense classification and localization. Recent trend for one-stage detectors is to introduce an individual prediction branch to estimate the quality of localization. This paper delves into the representations of the above three fundamental elements: quality estimation, classification and localization.
  arXiv  Detail & Related papers  (2020-06-08T07:24:33Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.