The field of computer vision and intelligent transportation systems is moving towards more accurate and realistic modeling of complex scenarios. Researchers are developing innovative methods to improve camera calibration, traffic simulation, and autonomous driving training. A key direction is the integration of generic and parametric models for camera calibration, which enhances accuracy and mitigates overfitting. Another significant trend is the use of noise-aware learning strategies and generative models to simulate realistic traffic phenomena, such as phantom traffic jams. The development of differentiable physics-based camera simulators and scene reconstruction techniques is also advancing the field of embodied AI and robotics. Noteworthy papers include:

• A paper that proposes a generic-parametric hybrid calibration method, which consistently excels across various lens types and noise contamination.
• A paper that introduces a noise-aware generative microscopic traffic simulation, which outperforms traditional baselines in realism and benefits from explicitly engaging with data imperfection.
• A paper that presents a differentiable physics-based camera simulator, which enables gradient-based optimization in visual perception pipelines and enhances robotic perception performance.