no code implementations • 16 Nov 2023 • Yangzheng Wu, Michael Greenspan
This paper addresses the simulation-to-real domain gap in 6DoF PE, and proposes a novel self-supervised keypoint radial voting-based 6DoF PE framework, effectively narrowing this gap using a learnable kernel in RKHS.
1 code implementation • 3 Sep 2023 • Mohsen Zand, Ali Etemad, Michael Greenspan
We use normalizing flows to parameterize the noisy data at any arbitrary step of the diffusion process and utilize it as the prior in the reverse diffusion process.
1 code implementation • 31 Aug 2023 • Mohsen Zand, Ali Etemad, Michael Greenspan
Our experiments on two challenging benchmark datasets, CMU Mocap and Human3. 6M, demonstrate that our proposed method is able to effectively model the sequence information for motion prediction and outperform other techniques to set a new state-of-the-art.
1 code implementation • 15 Aug 2023 • Yangzheng Wu, Michael Greenspan
We address the problem of keypoint selection, and find that the performance of 6DoF pose estimation methods can be improved when pre-defined keypoint locations are learned, rather than being heuristically selected as has been the standard approach.
no code implementations • 7 Jul 2023 • Haleh Damirchi, Michael Greenspan, Ali Etemad
Quantitative results demonstrate the superiority of our proposed model over the current state-of-the-art, which consistently achieves the lowest error for 3 time horizons of 0. 5, 1. 0 and 1. 5 seconds.
1 code implementation • 26 Jun 2023 • Mahdiyar Molahasani, Ali Etemad, Michael Greenspan
A continual learning solution is proposed to address the out-of-distribution generalization problem for pedestrian detection.
no code implementations • 23 Jun 2023 • Mahdiyar Molahasani, Michael Greenspan, Ali Etemad
Next, we assert that by treating the learning of the Head and Tail as two separate and sequential steps, Continual Learning (CL) methods can effectively update the weights of the learner to learn the Tail without forgetting the Head.
no code implementations • 16 May 2023 • Andrew Farley, Mohsen Zand, Michael Greenspan
We propose a method that augments a simulated dataset using diffusion models to improve the performance of pedestrian detection in real-world data.
1 code implementation • 18 Oct 2022 • Mo Yu, Yi Gu, Xiaoxiao Guo, Yufei Feng, Xiaodan Zhu, Michael Greenspan, Murray Campbell, Chuang Gan
Hence, in order to achieve higher performance on our tasks, models need to effectively utilize such functional knowledge to infer the outcomes of actions, rather than relying solely on memorizing facts.
1 code implementation • 14 Oct 2022 • Yangzheng Wu, Alireza Javaheri, Mohsen Zand, Michael Greenspan
We propose a novel keypoint voting 6DoF object pose estimation method, which takes pure unordered point cloud geometry as input without RGB information.
1 code implementation • 14 Jul 2022 • Mohsen Zand, Ali Etemad, Michael Greenspan
We present ObjectBox, a novel single-stage anchor-free and highly generalizable object detection approach.
1 code implementation • 9 Mar 2022 • Yufei Feng, Xiaoyu Yang, Xiaodan Zhu, Michael Greenspan
We introduce a neuro-symbolic natural logic framework based on reinforcement learning with introspective revision.
1 code implementation • 21 Feb 2022 • Mohsen Zand, Haleh Damirchi, Andrew Farley, Mahdiyar Molahasani, Michael Greenspan, Ali Etemad
As the detection and localization tasks are well-correlated and can be jointly tackled, our model benefits from a multitask solution by learning multiscale representations of encoded crowd images, and subsequently fusing them.
no code implementations • 12 Jun 2021 • Joy Mazumder, Mohsen Zand, Michael Greenspan
Applying our method can also improve the pose estimation average precision results of Op-Net by 6. 06% on average.
no code implementations • 24 Apr 2021 • Mohsen Zand, Ali Etemad, Michael Greenspan
A novel object detection method is presented that handles freely rotated objects of arbitrary sizes, including tiny objects as small as $2\times 2$ pixels.
1 code implementation • 9 Apr 2021 • Mohsen Zand, Ali Etemad, Michael Greenspan
We specifically propose to use conditional priors to factorize the latent space for the time dependent modeling.
1 code implementation • ICCV 2021 • Hardik Uppal, Alireza Sepas-Moghaddam, Michael Greenspan, Ali Etemad
Moreover, face recognition experiments demonstrate that our hallucinated depth along with the input RGB images boosts performance across various architectures when compared to a single RGB modality by average values of +1. 2%, +2. 6%, and +2. 6% for IIIT-D, EURECOM, and LFW datasets respectively.
1 code implementation • 6 Apr 2021 • Yangzheng Wu, Mohsen Zand, Ali Etemad, Michael Greenspan
We propose a novel keypoint voting scheme based on intersecting spheres, that is more accurate than existing schemes and allows for fewer, more disperse keypoints.
Ranked #1 on 6D Pose Estimation using RGBD on YCB-Video (ADDS AUC metric)
no code implementations • 22 Feb 2021 • Ghani O. Lawal, Michael Greenspan
A novel user friendly method is proposed for calibrating a procam system from a single pose of a planar chessboard target.
1 code implementation • 3 Jan 2021 • Hardik Uppal, Alireza Sepas-Moghaddam, Michael Greenspan, Ali Etemad
Our novel attention mechanism directs the deep network "where to look" for visual features in the RGB image by focusing the attention of the network using depth features extracted by a Convolution Neural Network (CNN).
1 code implementation • COLING 2020 • Yufei Feng, Zi'ou Zheng, Quan Liu, Michael Greenspan, Xiaodan Zhu
We explore end-to-end trained differentiable models that integrate natural logic with neural networks, aiming to keep the backbone of natural language reasoning based on the natural logic formalism while introducing subsymbolic vector representations and neural components.
no code implementations • 19 Oct 2020 • Mo Yu, Xiaoxiao Guo, Yufei Feng, Xiaodan Zhu, Michael Greenspan, Murray Campbell
Commonsense reasoning simulates the human ability to make presumptions about our physical world, and it is an indispensable cornerstone in building general AI systems.
no code implementations • 6 Apr 2020 • Yufei Feng, Mo Yu, Wenhan Xiong, Xiaoxiao Guo, Jun-Jie Huang, Shiyu Chang, Murray Campbell, Michael Greenspan, Xiaodan Zhu
We propose the new problem of learning to recover reasoning chains from weakly supervised signals, i. e., the question-answer pairs.
1 code implementation • 29 Feb 2020 • Hardik Uppal, Alireza Sepas-Moghaddam, Michael Greenspan, Ali Etemad
A novel attention aware method is proposed to fuse two image modalities, RGB and depth, for enhanced RGB-D facial recognition.