3D City mapping and deep learning from LiDAR point clouds by Prof Ajmal Mian – UNSW.ai Workshop

hqdefault 10

UNSW AI Institute Workshop – Imaging, Sensing and Data Informatics with AI ( Part 1)
3D point clouds and meshes are an important data source for vision tasks. In this talk, I will introduce our sweet of novel convolutional kernels designed for deep learning over 3D meshes. These kernels include vertex2vertex, vertex2facet, facet2vertex, and facet2facet convolutions and are available on GitHub as the Picasso library. Picasso also contains the first CUDA-based on-the-fly mesh decimation algorithm to facilitate hierarchical deep learning and integrates with PyTorch and TensorFlow to facilitate the design of novel network architectures. Whereas mesh convolutions are important for precise 3D object shape analysis, many vision tasks such as object detection, tracking, semantic segmentation and localization can be performed directly on the raw point clouds. For this purpose, transformer architectures have recently gained popularity. In the second part of the talk, I will discuss our research on self-supervised learning for multiple object tracking, object classification and self-localization. I will also present our work on LiDAR point cloud based semantic segmentation and simultaneous detection and tracking of multiple objects. Finally, I will discuss the challenges faced by the transformer architecture when learning from outdoor LiDAR data.

Ajmal Mian is a Professor of Computer Science at The University of Western Australia. He is the recipient of three prestigious national level fellowships from the Australian Research Council (ARC) including the recent Future Fellowship award 2022. Prof Ajmal is a Fellow of the International Association for Pattern Recognition, Distinguished Speaker of the Association for Computing Machinery and President of the Australian Pattern Recognition Society. He received the West Australian Early Career Scientist of the Year Award 2012 and the HBF Mid-Career Scientist of the Year Award 2022. Ajmal Mian has secured research funding from the ARC, the National Health and Medical Research Council of Australia, US Department of Defence DARPA, and the Australian Department of Defence. He is a Senior Editor for IEEE Transactions on Neural Networks & Learning Systems and an Associate Editor for IEEE Transactions on Image Processing and the Pattern Recognition journal. He served as a General Chair of the International Conference on Digital Image Computing Techniques & Applications (DICTA 2019) and the Asian Conference on Computer Vision (ACCV 2018). He also served as an Area Chair of ECCV 2022, CVPR 2022, ACM Multimedia 2020, WACV 2019, WACV 2018, ICPR 2016 and ACCV 2014. He has supervised 22 PhD students to completion and has published over 240 scientific papers. His research areas include computer vision, deep learning, 3D point cloud analysis, facial recognition, human action recognition and video analysis.

By: UNSW Community
Title: 3D City mapping and deep learning from LiDAR point clouds by Prof Ajmal Mian – UNSW.ai Workshop
Sourced From: www.youtube.com/watch?v=Azy6DGmqeN4

Follow us on FB: https://www.facebook.com/UXClub.net/

Did you miss our previous article…
https://www.uxclub.net/webinar-25-october-2022-advanced-rsi-trading-strategies/

Leave a Reply