Current state-of-the-art object detection algorithms still suffer the problem of imbalanced distribution of training data over object classes and background. Recent work introduced a new loss function called focal loss to mitigate this problem, but at the cost of an additional hyperparameter. Manually tuning this hyperparameter for each training task is highly time-consuming.With automated focal loss we introduce a new loss function which substitutes this hyperparameter by a parameter that is automatically adapted during the training progress and controls the amount of focusing on hard training examples. We show on the COCO benchmark that this leads to an up to 30 % faster training convergence. We further introduced a focal regression loss which on the more challenging task of 3D vehicle detection outperforms other loss functions by up to 1.8 AOS and can be used as a value range independent metric for regression.
Zusammenfassung. Es wird gezeigt, daf aIle primitiven pythagoreischen Tripel ganzer Zahlen aus (0,1,1) durch mehrfaches Anwenden einer linearen Transformation A und eventuelle Anderungen des Vorzeichens entstehen.In [1] zeigt Ryden, daf unendlich viele pythagoreische Tripel der Form (a,a+l,c) aus (3,4,5) verrnoge der Rekursion an+! = 3an + 2cn + 1 Cn+! = 4an + 3cn + 2 entstehen. In [2] weist Hering nach, daf man auf diese Art aile pythagoreischen Tripel der Form (a, a + 1, c) erhalt, Diese Rekursion liiBt sich noch umformen zu an+l = 6an -an-l + 2 Cn+l = 6cn -Cn-l , denn 6an -an-l + 2 = 3(6an_l -an-2 + 2) + 2(6c n_l -C n -2) + 1 und 6cn -Cn-l = 4(6an-l -an-2 + 2) + 3(6an_ l -Cn-2 ) + 2 . Mit b., = an + 1 kann man die eingangs angegebene Rekursion auch in der Form an+l = 2an + bn + 2cn bn+1 = an + 2bn + 2cn Cn+l = 2an + 2bn + 3cn
Tremendous progress in deep learning over the last years has led towards a future with autonomous vehicles on our roads. Nevertheless, the performance of their perception systems is strongly dependent on the quality of the utilized training data. As these usually only cover a fraction of all object classes an autonomous driving system will face, such systems struggle with handling the unexpected. In order to safely operate on public roads, the identification of objects from unknown classes remains a crucial task. In this paper, we propose a novel pipeline to detect unknown objects. Instead of focusing on a single sensor modality, we make use of lidar and camera data by combining state-of-the art detection models in a sequential manner. We evaluate our approach on the Waymo Open Perception Dataset and point out current research gaps in anomaly detection.
Autonomous driving is a key technology towards a brighter, more sustainable future. To enable such a future, it is necessary to utilize autonomous vehicles in shared mobility models. However, to evaluate, whether two or more route requests have the potential for a shared ride, is a computeintensive task, if done by rerouting. In this work, we propose the Dynamic Longest Common Subsequences algorithm for fast and cost-efficient comparison of two routes for their compatibility, dynamically only incorporating parts of the routes which are suited for a shared trip. Based on this, one can also estimate, how many autonomous vehicles might be necessary to fulfill the local mobility demands. This can help providers to estimate the necessary fleet sizes, policymakers to better understand mobility patterns and cities to scale necessary infrastructure.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.