no code implementations • 16 Apr 2024 • Hieu Le, Zhenhua He, Mai Le, Dhruva K. Chakravorty, Lisa M. Perez, Akhil Chilumuru, Yan Yao, Jiefu Chen
The discoveries in this paper show that Intelligence Processing Units (IPUs) offer a viable accelerator alternative to GPUs for machine learning (ML) applications within the fields of materials science and battery research.
no code implementations • 25 Mar 2024 • Nikita Durasov, Doruk Oner, Jonathan Donier, Hieu Le, Pascal Fua
Turning pass-through network architectures into iterative ones, which use their own output as input, is a well-known approach for boosting performance.
no code implementations • 22 Sep 2023 • Jingyi Xu, Hieu Le, Dimitris Samaras
Thus, we propose zero-shot object counting (ZSC), a new setting where only the class name is available during test time.
1 code implementation • 18 Sep 2023 • An Dinh Vuong, Minh Nhat Vu, Hieu Le, Baoru Huang, Binh Huynh, Thieu Vo, Andreas Kugi, Anh Nguyen
Foundation models such as ChatGPT have made significant strides in robotic tasks due to their universal representation of real-world domains.
no code implementations • 16 Jul 2023 • Hieu Le, Nicolas Talabot, Jiancheng Yang, Pascal Fua
Further, we show that our proposed method can be used to simulate various ways a hand can interact with an arbitrary object.
no code implementations • 15 Jul 2023 • Jingyi Xu, Hieu Le, Dimitris Samaras
In this paper, we point out that the task of counting objects of interest when there are multiple object classes in the image (namely, multi-class object counting) is particularly challenging for current object counting models.
1 code implementation • 9 Jul 2023 • Hieu Le, Hernan Santos, Jian Tao
Our model achieves a compression ratio of 140 on several benchmark data sets without compromising the reconstruction quality.
no code implementations • CVPR 2023 • Jingyi Xu, Hieu Le, Dimitris Samaras
To mitigate this issue, we propose a novel variational autoencoder (VAE) based data generation model, which is capable of generating data with increased crop-related diversity.
1 code implementation • CVPR 2023 • Jingyi Xu, Hieu Le, Vu Nguyen, Viresh Ranjan, Dimitris Samaras
By applying this model to all the candidate patches, we can select the most suitable patches as exemplars for counting.
Ranked #4 on Zero-Shot Counting on FSC147
no code implementations • 21 Nov 2022 • Nikita Durasov, Nik Dorndorf, Hieu Le, Pascal Fua
Sampling-free approaches can be faster but suffer from other drawbacks, such as lower reliability of uncertainty estimates, difficulty of use, and limited applicability to different types of tasks and data.
1 code implementation • 7 Oct 2022 • Tien-Phat Nguyen, Trong-Thang Pham, Tri Nguyen, Hieu Le, Dung Nguyen, Hau Lam, Phong Nguyen, Jennifer Fowler, Minh-Triet Tran, Ngan Le
The transformer expanding path models the temporal coherency between embryo images to ensure monotonic non-decreasing constraint and is optimized by a segmentation head.
no code implementations • 16 May 2022 • Hieu Le, Taufiq Daryanto, Fabian Zhafransyah, Derry Wijaya, Elizabeth Coppock, Sang Chin
This paper focuses on a referring expression generation (REG) task in which the aim is to pick out an object in a complex visual scene.
1 code implementation • CVPR 2022 • Jingyi Xu, Hieu Le
To mitigate this issue, we propose to generate visual samples based on semantic embeddings using a conditional variational autoencoder (CVAE) model.
1 code implementation • 25 Feb 2022 • Hieu Le, Salma Elmalaki, Athina Markopoulou, Zubair Shafiq
AutoFR is effective: it generates filter rules that can block 86% of the ads, as compared to 87% by EasyList, while achieving comparable visual breakage.
no code implementations • 5 Jan 2022 • Hieu Le, Hans Walker, Dung Tran, Peter Chin
Although deep neural networks have achieved great performance on classification tasks, recent studies showed that well trained networks can be fooled by adding subtle noises.
no code implementations • 29 Jul 2021 • Shilin Hu, Hieu Le, Dimitris Samaras
The current video shadow detection method achieves this goal via co-attention, which mostly exploits information that is temporally coherent but is not robust in detecting moving shadows and small shadow regions.
2 code implementations • 23 Dec 2020 • Hieu Le, Dimitris Samaras
Inspired by physical models of shadow formation, we use a linear illumination transformation to model the shadow effects in the image that allows the shadow image to be expressed as a combination of the shadow-free image, the shadow parameters, and a matte layer.
Ranked #2 on Shadow Removal on Adjusted ISTD
no code implementations • ICCV 2021 • Jingyi Xu, Hieu Le, Mingzhen Huang, ShahRukh Athar, Dimitris Samaras
We assume that the distribution of intra-class variance generalizes across the base class and the novel class.
Ranked #14 on Few-Shot Image Classification on CUB 200 5-way 5-shot
no code implementations • ECCV 2020 • Hieu Le, Dimitris Samaras
Our method achieves competitive shadow removal results compared to state-of-the-art methods that are trained with fully paired shadow and shadow-free images.
Ranked #4 on Shadow Removal on Adjusted ISTD
3 code implementations • ICCV 2019 • Hieu Le, Dimitris Samaras
Training our model on this new augmented ISTD dataset further lowers the RMSE on the shadow area to 7. 4.
Ranked #3 on Shadow Removal on Adjusted ISTD
no code implementations • 8 May 2019 • Hieu Le, Bento Gonçalves, Dimitris Samaras, Heather Lynch
This segmentation network is trained with a specific loss function, based on the average activation, to effectively learn from the data with the weakly-annotated labels.
no code implementations • ECCV 2018 • Viresh Ranjan, Hieu Le, Minh Hoai
In this work, we tackle the problem of crowd counting in images.
Ranked #10 on Crowd Counting on UCF CC 50
1 code implementation • ECCV 2018 • Hieu Le, Tomas F. Yago Vicente, Vu Nguyen, Minh Hoai, Dimitris Samaras
The A-Net modifies the original training images constrained by a simplified physical shadow model and is focused on fooling the D-Net's shadow predictions.
Ranked #4 on Shadow Detection on SBU
no code implementations • 31 Mar 2017 • Hieu Le, Vu Nguyen, Chen-Ping Yu, Dimitris Samaras
This paper proposes a geodesic-distance-based feature that encodes global information for improved video segmentation algorithms.
Ranked #1 on Video Segmentation on SegTrack v2
no code implementations • 10 Dec 2016 • Hieu Le, Chen-Ping Yu, Gregory Zelinsky, Dimitris Samaras
Co-localization is the problem of localizing objects of the same class using only the set of images that contain them.
Ranked #1 on Object Localization on PASCAL VOC 2012
no code implementations • ICCV 2015 • Chen-Ping Yu, Hieu Le, Gregory Zelinsky, Dimitris Samaras
Video segmentation is the task of grouping similar pixels in the spatio-temporal domain, and has become an important preprocessing step for subsequent video analysis.