LATIDIA · Robótica

TRIM: Monitoreo temporal en tiempo de ejecución de la detección de objetos 3D a través de estadísticas de activación

arXiv: 2610.38613v1Tipo de anuncio: cross Resumen: La percepción confiable es esencial para los sistemas de transporte inteligentes, sin embargo, las redes neuronales profundas pueden pasar por alto silenciosamente objetos críticos para la seguridad mientras continúan produciendo aplicaciones

La noticia

arXiv:2610.38613v1 Announce Type: cross Abstract: Reliable perception is essential to intelligent transportation systems, yet deep neural networks can silently miss safety-critical objects while continuing to produce apparently plausible outputs. Existing run-time monitors for 3D object detection generally assess each frame independently and therefore cannot exploit the temporal evolution of occlusion, point sparsity, and scene complexity. We present TRIM, a lightweight temporal monitor that detects missed-object failures without modifying or retraining the underlying detector. TRIM summarises intermediate activations using compact channel-wise statistics and processes short sequences of these summaries with a temporal convolutional network. Experiments on NuScenes show that temporal monitoring improves AUROC from 0.686 to 0.704 and reduces the false