PerceptionMetrics: GUI extension and support for standard datasets and models

Perception metrics is a unified toolkit for evaluating perception models across datasets and frameworks. Its foundation was built through prior GSoC efforts designed to unify and formalized in a stable release published in Sensors. This platform was recently modernized to support LiDAR, images segmentation, and object detection. This proposal focuses on scaling PerceptionMetrics from off-road focused workflows to industry-standard benchmarking. The core objectives are: • Integrate and prioritize standard image and Lidar datasets. • Extend GUI support to segmentation visualization (image and Lidar). • Deliver comprehensive tutorials/documentations for practical usage. • Strengthen reliability through expanded tests and CI-friendly validation.