ModelNet40
收藏资源简介:
ModelNet40数据集是由视觉智能卓越中心创建的,用于3D点云增量学习的基准数据集。该数据集包含40个类别的3D模型,用于评估不同点云分类方法的性能。创建过程中,数据集被设计为不包含先前任务的数据,以模拟实际应用中数据不断增加的情况。该数据集主要应用于3D计算机视觉领域,旨在解决模型在新数据到来时如何有效学习而不遗忘旧知识的问题。
The ModelNet40 dataset, developed by the Centre of Excellence for Visual Intelligence, is a benchmark dataset for 3D point cloud incremental learning. It contains 3D models across 40 categories, and is used to evaluate the performance of various point cloud classification methods. During its development, the dataset was designed to exclude data from prior tasks, simulating the scenario of continuously increasing data in real-world applications. Primarily applied in the field of 3D computer vision, this dataset aims to address the challenge of how models can effectively learn new knowledge without forgetting old information when new data arrives.

- 1PointCLIMB: An Exemplar-Free Point Cloud Class Incremental Benchmark视觉智能卓越中心(CEVI) · 2023年



