Mathieu Orzalesi

Monocular 3D object detection (Mono3D) is essential for autonomous driving and augmented reality, yet its performance degrades significantly at night due to the scarcity of annotated nighttime data. In this paper, we investigate the use of style transfer for nighttime data augmentation and evaluate its effect on individual components of 3D detection. Using CycleGAN, we generated synthetic night images from daytime scenes in the nuScenes dataset and trained a modular Mono3D detector under different configurations. Our results show that training solely on style-transferred images improves certain metrics, such as AP@0.95 (from 0.0299 to 0.0778, a 160% increase) and depth error (11% reduction), compared to daytime-only baselines. However, performance on orientation and dimension estimation deteriorates. When real nighttime data is included, style transfer provides complementary benefits: for cars, depth error decreases from 0.0414 to 0.021, and AP@0.95 remains stable at 0.66; for pedestrians, AP@0.95 improves by 13% (0.297 to 0.336) with a 35% reduction in depth error. Cyclist detection remains unreliable due to limited samples. We conclude that style transfer cannot replace authentic nighttime data, but when combined with it, it reduces false positives and improves depth estimation, leading to more robust detection under low-light conditions. This study highlights both the potential and the limitations of style transfer for augmenting Mono3D training, and it points to future research on more advanced generative models and broader object categories.

Multi-axis computer numerical control machines are increasingly present in the industry, as both production performance and the complexity of operational tasks continue to rise. In this context, the machine selection process remains a perpetual challenge due to the multitude of possible machine configurations. Machines are generally used to perform different tasks, which explains why selection methods remain to be developed with a view to optimizing machine configuration for a specific task while minimizing machine cost. Special-purpose machines are usually selected off-the-shelf, designed based on experience, or focused on overall performance, often resulting in inadequate performance for the required task. With the aim to help the designer in the architecture selection process at the pre-study stage, a new task-based optimization method is proposed in this article, allowing to identify both optimal machine geometric parameters and joint tolerances, while minimizing the cost of key machine components. Task specifications, represented by the spatio-temporal tool trajectory, are defined as optimization constraints. The proposed method is applied to the selection of a five-axis machine architecture in the context of additive manufacturing for medical orthoses. The step-by-step optimization process, along with the results on costs and optimal parameters, is presented and discussed.