Abstract: Object detection from point clouds is a fundamental task for 3D scene understanding and has a wide range of applications in the field of multimedia data processing and analysis, such as ...

BlendCLIP is a multimodal pretraining framework that bridges this synthetic-to-real gap by strategically combining the strengths of both domains. It introduces a curriculum-based data mixing strategy ...

Abstract: Monocular 3D object detection (Mono3D) holds noteworthy promise for autonomous driving applications owing to the cost-effectiveness and rich visual context of monocular camera sensors.

4D millimeter-wave radar has emerged as a promising sensor for autonomous driving, but effective 3D object detection from both 4D radar and monocular images remains a challenge. Existing fusion ...