Monocular temporal 3D detection aims to detect objects in 3D, given a monocular video. Query-based 3D detectors unify detection and cross-view association, but their learnable queries fit the spatial ...

Real-time monocular 3D object detection remains challenging due to severe depth ambiguity, viewpoint shifts, and the high computational cost of 3D reasoning. Existing approaches either rely on LiDAR o...

Diffusion Transformers (DiTs) have emerged as a dominant architecture for high-quality text-to-image generation, yet their scale poses challenges for efficient deployment. While truncated singular val...

PhD Student @ Visual Inference Lab (@visinf.bsky.social), TU Darmstadt

Knowing where and how fast objects are moving within a scene is important across various domains. Usually, cameras are used to capture the data necessary for this task, but adding cameras often raises...

Egocentric action recognition enables robots to facilitate human-robot interactions and monitor task progress. Existing methods often rely solely on RGB videos, although additional modalities, such as...

Unsupervised object-centric learning (OCL) decomposes visual scenes into distinct entities. Slot attention is a popular approach that represents individual objects as latent vectors, called slots. Cur...

Recent advances in semantic correspondence rely on dual-encoder architectures, combining DINOv2 with diffusion backbones. While accurate, these billion-parameter models generalize poorly beyond traini...

Video panoptic segmentation (VPS) aims to jointly detect, segment, and track all objects while partitioning the video into semantically consistent regions. We introduce the task setting of unsupervise...

Multimodal Knowledge Distillation for Egocentric Action Recognition Robust to Missing Modalities.

Kyutai and ELLIS Tübingen announce the official launch of KE:SAI – Kyutai ELLIS Scalable Autonomous Intelligence, a non-profit open science lab for physical AI.

We present ReCoVEr - A training strategy for removing the cost volume from an optical flow estimator during training.