MBZUAI Research Talks | 3D POINT CLOUD ANALYSIS AND RGBD SALIENCY DETECTION | Dr. Saeed Anwar
Published at : October 01, 2021
We will be analyzing 3D models in this seminar. Firstly, we propose a plug-and-plug module for point clouds. Secondly, we present an uncertainty-inspired RGB-D saliency detection method.
Recently, many points cloud networks have been invented for visual analysis with the help of the deep learning paradigm. However, there is great potential for advancing these networks since the given information of point cloud data has not been fully exploited. To improve the effectiveness of existing networks in analyzing point cloud data, we propose a plug-and-play module, PnP-3D, aiming to refine the fundamental point cloud feature representations by involving more local context and global bilinear response from explicit 3D space and implicit features. To thoroughly evaluate our approach, we conduct experiments on three standard tasks, including classification, semantic segmentation, and object detection, selecting three state-of-the-art networks from each task for evaluation. Serving as a plug-and-play module, PnP-3D can significantly boost the performances of established networks. In addition to achieving state-of-the-art results on four widely used point cloud benchmarks.
Similarly, for RGB-D saliency detection, we propose the first framework (UCNet) to employ uncertainty by learning from the data labeling process. Existing RGB-D saliency detection methods treat the saliency detection task as a point estimation problem and produce a single saliency map following a deterministic learning pipeline. Inspired by the saliency data labeling process, we propose a probabilistic RGB-D saliency detection network via conditional variational autoencoders to model human annotation uncertainty and generate multiple saliency maps for each input image by sampling in the latent space. We can generate an accurate saliency map based on these multiple predictions with the proposed saliency consensus process.
Recently, many points cloud networks have been invented for visual analysis with the help of the deep learning paradigm. However, there is great potential for advancing these networks since the given information of point cloud data has not been fully exploited. To improve the effectiveness of existing networks in analyzing point cloud data, we propose a plug-and-play module, PnP-3D, aiming to refine the fundamental point cloud feature representations by involving more local context and global bilinear response from explicit 3D space and implicit features. To thoroughly evaluate our approach, we conduct experiments on three standard tasks, including classification, semantic segmentation, and object detection, selecting three state-of-the-art networks from each task for evaluation. Serving as a plug-and-play module, PnP-3D can significantly boost the performances of established networks. In addition to achieving state-of-the-art results on four widely used point cloud benchmarks.
Similarly, for RGB-D saliency detection, we propose the first framework (UCNet) to employ uncertainty by learning from the data labeling process. Existing RGB-D saliency detection methods treat the saliency detection task as a point estimation problem and produce a single saliency map following a deterministic learning pipeline. Inspired by the saliency data labeling process, we propose a probabilistic RGB-D saliency detection network via conditional variational autoencoders to model human annotation uncertainty and generate multiple saliency maps for each input image by sampling in the latent space. We can generate an accurate saliency map based on these multiple predictions with the proposed saliency consensus process.
MBZUAIResearchTalks