Sparse keypoint matching is crucial for 3D vision tasks, yet current keypoint detectors often produce spatially inaccurate matches. Existing refinement methods mitigate this issue through alignment of matched keypoint locations, but they are typically detector-specific, requiring retraining for each keypoint detector. We introduce XRefine, a novel, detector-agnostic approach for sub-pixel keypoint refinement that operates solely on image patches centered at matched keypoints. Our cross-attention-based architecture learns to predict refined keypoint coordinates without relying on internal detector representations, enabling generalization across detectors. Furthermore, XRefine can be extended to handle multi-view feature tracks. Experiments on MegaDepth, KITTI, and ScanNet demonstrate that the approach consistently improves geometric estimation accuracy, achieving superior performance compared to existing refinement methods while maintaining runtime efficiency. Our code and trained models can be found at https://github.com/boschresearch/xrefine.

Accurate intrinsic calibration is essential for camera-based 3D perception, yet, it typically requires targets of well-known geometry. Here, we propose a camera self-calibration approach that infers camera intrinsics during application, from monocular videos in the wild. We propose to explicitly model projection functions and multi-view geometry, while leveraging the capabilities of deep neural networks for feature extraction and matching. To achieve this, we build upon recent research on integrating bundle adjustment into deep learning models, and introduce a self-calibrating bundle adjustment layer. The self-calibrating bundle adjustment layer optimizes camera intrinsics through classical Gauß-Newton steps and can be adapted to different camera models without re-training. As a specific realization, we implemented this layer within the deep visual SLAM system DROID-SLAM, and show that the resulting model, DroidCalib, yields state-of-the-art calibration accuracy across multiple public datasets. Our results suggest that the model generalizes to unseen environments and different camera models, including significant lens distortion. Thereby, the approach enables performing 3D perception tasks without prior knowledge about the camera. Code is available at https://github.com/boschresearch/droidcalib.

Most approaches to camera calibration rely on calibration targets of well-known geometry. During data acquisition, calibration target and camera system are typically moved w.r.t. each other, to allow image coverage and perspective versatility. We show that moving the target can lead to small temporary deformations of the target, which can introduce significant errors into the calibration result. While static inaccuracies of calibration targets have been addressed in previous works, to our knowledge, none of the existing approaches can capture time-varying, dynamic deformations. To achieve high-accuracy calibrations despite moving the target, we propose a way to explicitly model dynamic target deformations in camera calibration. This is achieved by using a low-dimensional deformation model with only few parameters per image, which can be optimized jointly with target poses and intrinsics. We demonstrate the effectiveness of modeling dynamic deformations using different calibration targets and show its significance in a structure-from-motion application.

Epileptic seizures are characterized by abnormal and excessive neural activity, where cortical network dynamics seem to become unstable. However, most of the time, during seizure-free periods, cortex of epilepsy patients shows perfectly stable dynamics. This raises the question of how recurring instability can arise in the light of this stable default state. In this work, we examine two potential scenarios of seizure generation: (i) epileptic cortical areas might generally operate closer to instability, which would make epilepsy patients generally more susceptible to seizures, or (ii) epileptic cortical areas might drift systematically towards instability before seizure onset. We analyzed single-unit spike recordings from both the epileptogenic (focal) and the nonfocal cortical hemispheres of 20 epilepsy patients. We quantified the distance to instability in the framework of criticality, using a novel estimator, which enables an unbiased inference from a small set of recorded neurons. Surprisingly, we found no evidence for either scenario: Neither did focal areas generally operate closer to instability, nor were seizures preceded by a drift towards instability. In fact, our results from both pre-seizure and seizure-free intervals suggest that despite epilepsy, human cortex operates in the stable, slightly subcritical regime, just like cortex of other healthy mammalians.