Search Results for author: Bernard Ghanem

Found 145 papers, 61 papers with code

Certified Robustness in Federated Learning

1 code implementation6 Jun 2022 Motasem Alfarra, Juan C. Pérez, Egor Shulgin, Peter Richtárik, Bernard Ghanem

However, as in the single-node supervised learning setup, models trained in federated learning suffer from vulnerability to imperceptible input transformations known as adversarial attacks, questioning their deployment in security-related applications.

Federated Learning

Egocentric Video-Language Pretraining

1 code implementation3 Jun 2022 Kevin Qinghong Lin, Alex Jinpeng Wang, Mattia Soldan, Michael Wray, Rui Yan, Eric Zhongcong Xu, Difei Gao, RongCheng Tu, Wenzhe Zhao, Weijie Kong, Chengfei Cai, Hongfa Wang, Dima Damen, Bernard Ghanem, Wei Liu, Mike Zheng Shou

Video-Language Pretraining (VLP), aiming to learn transferable representation to advance a wide range of video-text downstream tasks, has recently received increasing attention.

Action Recognition Contrastive Learning +2

ETAD: A Unified Framework for Efficient Temporal Action Detection

1 code implementation14 May 2022 Shuming Liu, Mengmeng Xu, Chen Zhao, Xu Zhao, Bernard Ghanem

Interestingly, on ActivityNet-1. 3, it reaches 37. 78% average mAP, while only requiring 6 mins of training time and 1. 23 GB memory based on pre-extracted features.

Action Detection Video Understanding

SoccerNet-Tracking: Multiple Object Tracking Dataset and Benchmark in Soccer Videos

no code implementations14 Apr 2022 Anthony Cioppa, Silvio Giancola, Adrien Deliege, Le Kang, Xin Zhou, Zhiyu Cheng, Bernard Ghanem, Marc Van Droogenbroeck

Tracking objects in soccer videos is extremely important to gather both player and team statistics, whether it is to estimate the total distance run, the ball possession or the team formation.

Multiple Object Tracking

3DeformRS: Certifying Spatial Deformations on Point Clouds

1 code implementation CVPR 2022 Gabriel Pérez S., Juan C. Pérez, Motasem Alfarra, Silvio Giancola, Bernard Ghanem

In this work, we propose 3DeformRS, a method to certify the robustness of point cloud Deep Neural Networks (DNNs) against real-world deformations.

Autonomous Driving

When NAS Meets Trees: An Efficient Algorithm for Neural Architecture Search

no code implementations11 Apr 2022 Guocheng Qian, Xuanyang Zhang, Guohao Li, Chen Zhao, Yukang Chen, Xiangyu Zhang, Bernard Ghanem, Jian Sun

TNAS performs a modified bi-level Breadth-First Search in the proposed trees to discover a high-performance architecture.

Neural Architecture Search

Real-time Hyperspectral Imaging in Hardware via Trained Metasurface Encoders

1 code implementation CVPR 2022 Maksim Makarenko, Arturo Burguete-Lopez, Qizhou Wang, Fedor Getman, Silvio Giancola, Bernard Ghanem, Andrea Fratalocchi

Hyperspectral imaging has attracted significant attention to identify spectral signatures for image classification and automated pattern recognition in computer vision.

Image Classification Semantic Segmentation +1

End-to-End Active Speaker Detection

no code implementations27 Mar 2022 Juan Leon Alcazar, Moritz Cordes, Chen Zhao, Bernard Ghanem

Recent advances in the Active Speaker Detection (ASD) problem build upon a two-stage process: feature extraction and spatio-temporal context aggregation.

R-DFCIL: Relation-Guided Representation Learning for Data-Free Class Incremental Learning

no code implementations24 Mar 2022 Qiankun Gao, Chen Zhao, Bernard Ghanem, Jian Zhang

After RRL, the classification head is fine-tuned with global class-balanced classification loss to address the data imbalance issue as well as learn the decision boundary between new and previous classes.

class-incremental learning Incremental Learning +2

Learning Scene Flow in 3D Point Clouds with Noisy Pseudo Labels

no code implementations23 Mar 2022 Bing Li, Cheng Zheng, Guohao Li, Bernard Ghanem

To provide an alternative, we propose a novel approach that utilizes monocular RGB images and point clouds to generate pseudo scene flow labels for training scene flow networks.

Self-Supervised Learning

SegTAD: Precise Temporal Action Detection via Semantic Segmentation

no code implementations3 Mar 2022 Chen Zhao, Merey Ramazanova, Mengmeng Xu, Bernard Ghanem

To address these issues and precisely model temporal action detection, we formulate the task of temporal action detection in a novel perspective of semantic segmentation.

Action Detection object-detection +2

OWL (Observe, Watch, Listen): Localizing Actions in Egocentric Video via Audiovisual Temporal Context

no code implementations10 Feb 2022 Merey Ramazanova, Victor Escorcia, Fabian Caba Heilbron, Chen Zhao, Bernard Ghanem

In this work, we take a deep look into the effectiveness of audio in detecting actions in egocentric videos and introduce a simple-yet-effective approach via Observing, Watching, and Listening (OWL) to leverage audio-visual information and context for egocentric TAL.

Temporal Action Localization Temporal Localization

Towards Assessing and Characterizing the Semantic Robustness of Face Recognition

no code implementations10 Feb 2022 Juan C. Pérez, Motasem Alfarra, Ali Thabet, Pablo Arbeláez, Bernard Ghanem

We propose a methodology for assessing and characterizing the robustness of FRMs against semantic perturbations to their input.

Face Recognition

On the Robustness of Quality Measures for GANs

no code implementations31 Jan 2022 Motasem Alfarra, Juan C. Pérez, Anna Frühstück, Philip H. S. Torr, Peter Wonka, Bernard Ghanem

We show the vulnerability of both the generative model and the FID against additive perturbations in the latent space.

Voint Cloud: Multi-View Point Cloud Representation for 3D Understanding

no code implementations30 Nov 2021 Abdullah Hamdi, Silvio Giancola, Bernard Ghanem

This novel 3D Voint cloud representation combines the compactness of 3D point cloud representation with the natural view-awareness of multi-view representation.

3D Classification 3D Semantic Segmentation

ASSANet: An Anisotropic Separable Set Abstraction for Efficient Point Cloud Representation Learning

1 code implementation NeurIPS 2021 Guocheng Qian, Hasan Abed Al Kader Hammoud, Guohao Li, Ali Thabet, Bernard Ghanem

We then introduce a new Anisotropic Reduction function into our Separable SA module and propose an Anisotropic Separable SA (ASSA) module that substantially increases the network's accuracy.

3D Part Segmentation 3D Point Cloud Classification +2

Ego4D: Around the World in 3,000 Hours of Egocentric Video

1 code implementation CVPR 2022 Kristen Grauman, Andrew Westbury, Eugene Byrne, Zachary Chavis, Antonino Furnari, Rohit Girdhar, Jackson Hamburger, Hao Jiang, Miao Liu, Xingyu Liu, Miguel Martin, Tushar Nagarajan, Ilija Radosavovic, Santhosh Kumar Ramakrishnan, Fiona Ryan, Jayant Sharma, Michael Wray, Mengmeng Xu, Eric Zhongcong Xu, Chen Zhao, Siddhant Bansal, Dhruv Batra, Vincent Cartillier, Sean Crane, Tien Do, Morrie Doulaty, Akshay Erapalli, Christoph Feichtenhofer, Adriano Fragomeni, Qichen Fu, Abrham Gebreselasie, Cristina Gonzalez, James Hillis, Xuhua Huang, Yifei HUANG, Wenqi Jia, Weslie Khoo, Jachym Kolar, Satwik Kottur, Anurag Kumar, Federico Landini, Chao Li, Yanghao Li, Zhenqiang Li, Karttikeya Mangalam, Raghava Modhugu, Jonathan Munro, Tullie Murrell, Takumi Nishiyasu, Will Price, Paola Ruiz Puentes, Merey Ramazanova, Leda Sari, Kiran Somasundaram, Audrey Southerland, Yusuke Sugano, Ruijie Tao, Minh Vo, Yuchen Wang, Xindi Wu, Takuma Yagi, Ziwei Zhao, Yunyi Zhu, Pablo Arbelaez, David Crandall, Dima Damen, Giovanni Maria Farinella, Christian Fuegen, Bernard Ghanem, Vamsi Krishna Ithapu, C. V. Jawahar, Hanbyul Joo, Kris Kitani, Haizhou Li, Richard Newcombe, Aude Oliva, Hyun Soo Park, James M. Rehg, Yoichi Sato, Jianbo Shi, Mike Zheng Shou, Antonio Torralba, Lorenzo Torresani, Mingfei Yan, Jitendra Malik

We introduce Ego4D, a massive-scale egocentric video dataset and benchmark suite.

De-identification

Relation-aware Video Reading Comprehension for Temporal Language Grounding

1 code implementation EMNLP 2021 Jialin Gao, Xin Sun, Mengmeng Xu, Xi Zhou, Bernard Ghanem

Temporal language grounding in videos aims to localize the temporal span relevant to the given query sentence.

Reading Comprehension

Data-Dependent Randomized Smoothing

no code implementations29 Sep 2021 Motasem Alfarra, Adel Bibi, Philip Torr, Bernard Ghanem

In this work, we revisit Gaussian randomized smoothing and show that the variance of the Gaussian distribution can be optimized at each input so as to maximize the certification radius for the construction of the smooth classifier.

MovieCuts: A New Dataset and Benchmark for Cut Type Recognition

no code implementations12 Sep 2021 Alejandro Pardo, Fabian Caba Heilbron, Juan León Alcázar, Ali Thabet, Bernard Ghanem

Understanding movies and their structural patterns is a crucial task to decode the craft of video editing.

Check Your Other Door! Creating Backdoor Attacks in the Frequency Domain

no code implementations12 Sep 2021 Hasan Abed Al Kader Hammoud, Bernard Ghanem

Deep Neural Networks (DNNs) are ubiquitous and span a variety of applications ranging from image classification and facial recognition to medical image analysis and real-time object detection.

Backdoor Attack Image Classification +2

Learning to Cut by Watching Movies

1 code implementation ICCV 2021 Alejandro Pardo, Fabian Caba Heilbron, Juan León Alcázar, Ali Thabet, Bernard Ghanem

Video content creation keeps growing at an incredible pace; yet, creating engaging stories remains challenging and requires non-trivial video editing expertise.

Contrastive Learning

ANCER: Anisotropic Certification via Sample-wise Volume Maximization

1 code implementation9 Jul 2021 Francisco Eiras, Motasem Alfarra, M. Pawan Kumar, Philip H. S. Torr, Puneet K. Dokania, Bernard Ghanem, Adel Bibi

All prior art on randomized smoothing has focused on isotropic $\ell_p$ certification, which has the advantage of yielding certificates that can be easily compared among isotropic methods via $\ell_p$-norm radius.

DeformRS: Certifying Input Deformations with Randomized Smoothing

2 code implementations2 Jul 2021 Motasem Alfarra, Adel Bibi, Naeemullah Khan, Philip H. S. Torr, Bernard Ghanem

Deep neural networks are vulnerable to input deformations in the form of vector fields of pixel displacements and to other parameterized geometric deformations e. g. translations, rotations, etc.

Training Graph Neural Networks with 1000 Layers

4 code implementations14 Jun 2021 Guohao Li, Matthias Müller, Bernard Ghanem, Vladlen Koltun

Deep graph neural networks (GNNs) have achieved excellent results on various tasks on increasingly large graph datasets with millions of nodes and edges.

Graph Sampling Node Property Prediction

SCTN: Sparse Convolution-Transformer Network for Scene Flow Estimation

1 code implementation10 May 2021 Bing Li, Cheng Zheng, Silvio Giancola, Bernard Ghanem

We propose a novel scene flow estimation approach to capture and infer 3D motions from point clouds.

Scene Flow Estimation

Camera Calibration and Player Localization in SoccerNet-v2 and Investigation of their Representations for Action Spotting

no code implementations19 Apr 2021 Anthony Cioppa, Adrien Deliège, Floriane Magera, Silvio Giancola, Olivier Barnich, Bernard Ghanem, Marc Van Droogenbroeck

Specifically, we distill a powerful commercial calibration tool in a recent neural network architecture on the large-scale SoccerNet dataset, composed of untrimmed broadcast videos of 500 soccer games.

Action Spotting Camera Calibration +1

Temporally-Aware Feature Pooling for Action Spotting in Soccer Broadcasts

1 code implementation14 Apr 2021 Silvio Giancola, Bernard Ghanem

In this paper, we focus our analysis on action spotting in soccer broadcast, which consists in temporally localizing the main actions in a soccer game.

Action Spotting

AniGAN: Style-Guided Generative Adversarial Networks for Unsupervised Anime Face Generation

3 code implementations24 Feb 2021 Bing Li, Yuanlue Zhu, Yitong Wang, Chia-Wen Lin, Bernard Ghanem, Linlin Shen

Specifically, a new generator architecture is proposed to simultaneously transfer color/texture styles and transform local facial shapes into anime-like counterparts based on the style of a reference anime-face, while preserving the global structure of the source photo-face.

Face Generation Translation

High Quality Disparity Remapping With Two-Stage Warping

no code implementations ICCV 2021 Bing Li, Chia-Wen Lin, Cheng Zheng, Shan Liu, Junsong Yuan, Bernard Ghanem, C.-C. Jay Kuo

In the second stage, we derive another warping model to refine warping results in less important regions by eliminating serious distortions in shape, disparity and 3D structure.

DeeperGCN: Training Deeper GCNs with Generalized Aggregation Functions

no code implementations1 Jan 2021 Guohao Li, Chenxin Xiong, Ali Thabet, Bernard Ghanem

We add our generalized aggregation into a deep GCN framework and show it achieves state-of-the-art results in six benchmarks from OGB.

Point Cloud Classification Representation Learning

On the Decision Boundaries of Neural Networks. A Tropical Geometry Perspective

no code implementations1 Jan 2021 Motasem Alfarra, Adel Bibi, Hasan Abed Al Kader Hammoud, Mohamed Gaafar, Bernard Ghanem

This work tackles the problem of characterizing and understanding the decision boundaries of neural networks with piecewise linear non-linearity activations.

Network Pruning

SALA: Soft Assignment Local Aggregation for Parameter Efficient 3D Semantic Segmentation

no code implementations29 Dec 2020 Hani Itani, Silvio Giancola, Ali Thabet, Bernard Ghanem

Since it is learnable, this mapping is allowed to be different per layer instead of being applied uniformly throughout the depth of the network.

3D Semantic Segmentation

Data Dependent Randomized Smoothing

1 code implementation8 Dec 2020 Motasem Alfarra, Adel Bibi, Philip H. S. Torr, Bernard Ghanem

In this work, we revisit Gaussian randomized smoothing and show that the variance of the Gaussian distribution can be optimized at each input so as to maximize the certification radius for the construction of the smooth classifier.

SoccerNet-v2: A Dataset and Benchmarks for Holistic Understanding of Broadcast Soccer Videos

2 code implementations26 Nov 2020 Adrien Deliège, Anthony Cioppa, Silvio Giancola, Meisam J. Seikavandi, Jacob V. Dueholm, Kamal Nasrollahi, Bernard Ghanem, Thomas B. Moeslund, Marc Van Droogenbroeck

In this work, we propose SoccerNet-v2, a novel large-scale corpus of manual annotations for the SoccerNet video dataset, along with open challenges to encourage more research in soccer understanding and broadcast production.

Action Spotting Boundary Detection +4

MVTN: Multi-View Transformation Network for 3D Shape Recognition

1 code implementation ICCV 2021 Abdullah Hamdi, Silvio Giancola, Bernard Ghanem

MVTN exhibits clear performance gains in the tasks of 3D shape classification and 3D shape retrieval without the need for extra training supervision.

3D Classification 3D Object Retrieval +5

TSP: Temporally-Sensitive Pretraining of Video Encoders for Localization Tasks

1 code implementation23 Nov 2020 Humam Alwassel, Silvio Giancola, Bernard Ghanem

Extensive experiments show that using features trained with our novel pretraining strategy significantly improves the performance of recent state-of-the-art methods on three tasks: Temporal Action Localization, Action Proposal Generation, and Dense Video Captioning.

Action Classification Dense Video Captioning +2

VLG-Net: Video-Language Graph Matching Network for Video Grounding

1 code implementation19 Nov 2020 Mattia Soldan, Mengmeng Xu, Sisi Qu, Jesper Tegner, Bernard Ghanem

Grounding language queries in videos aims at identifying the time interval (or moment) semantically relevant to a language query.

Graph Matching Moment Retrieval +3

Robust Optimization as Data Augmentation for Large-scale Graphs

3 code implementations CVPR 2022 Kezhi Kong, Guohao Li, Mucong Ding, Zuxuan Wu, Chen Zhu, Bernard Ghanem, Gavin Taylor, Tom Goldstein

Data augmentation helps neural networks generalize better by enlarging the training set, but it remains an open question how to effectively augment graph data to enhance the performance of GNNs (Graph Neural Networks).

Data Augmentation Graph Classification +3

LC-NAS: Latency Constrained Neural Architecture Search for Point Cloud Networks

no code implementations24 Aug 2020 Guohao Li, Mengmeng Xu, Silvio Giancola, Ali Thabet, Bernard Ghanem

In this paper, we introduce a new NAS framework, dubbed LC-NAS, where we search for point cloud architectures that are constrained to a target latency.

Neural Architecture Search Point Cloud Classification +2

Learning Heat Diffusion for Network Alignment

no code implementations10 Jul 2020 Sisi Qu, Mengmeng Xu, Bernard Ghanem, Jesper Tegner

EDNA uses the diffusion signal as a proxy for computing node similarities between networks.

Network Moments: Extensions and Sparse-Smooth Attacks

no code implementations21 Jun 2020 Modar Alfadly, Adel Bibi, Emilio Botero, Salman AlSubaihi, Bernard Ghanem

This has incited research on the reaction of DNNs to noisy input, namely developing adversarial input attacks and strategies that lead to robust DNNs to these attacks.

Rethinking Clustering for Robustness

1 code implementation13 Jun 2020 Motasem Alfarra, Juan C. Pérez, Adel Bibi, Ali Thabet, Pablo Arbeláez, Bernard Ghanem

This paper studies how encouraging semantically-aligned features during deep neural network training can increase network robustness.

DeeperGCN: All You Need to Train Deeper GCNs

3 code implementations13 Jun 2020 Guohao Li, Chenxin Xiong, Ali Thabet, Bernard Ghanem

Graph Convolutional Networks (GCNs) have been drawing significant attention with the power of representation learning on graphs.

Graph Learning Graph Property Prediction +2

Adaptive Learning of the Optimal Batch Size of SGD

no code implementations3 May 2020 Motasem Alfarra, Slavomir Hanzely, Alyazeed Albasyoni, Bernard Ghanem, Peter Richtarik

Recent advances in the theoretical understanding of SGD led to a formula for the optimal batch size minimizing the number of effective data passes, i. e., the number of iterations times the batch size.

On the Decision Boundaries of Neural Networks: A Tropical Geometry Perspective

no code implementations20 Feb 2020 Motasem Alfarra, Adel Bibi, Hasan Hammoud, Mohamed Gaafar, Bernard Ghanem

Our main finding is that the decision boundaries are a subset of a tropical hypersurface, which is intimately related to a polytope formed by the convex hull of two zonotopes.

Network Pruning

RGB-based Semantic Segmentation Using Self-Supervised Depth Pre-Training

no code implementations6 Feb 2020 Jean Lahoud, Bernard Ghanem

These labels, denoted by HN-labels, represent different height and normal patches, which allow mining of local semantic information that is useful in the task of semantic RGB segmentation.

Semantic Segmentation

Analytical Moment Regularizer for Training Robust Networks

no code implementations ICLR 2020 Modar Alfadly, Adel Bibi, Muhammed Kocabas, Bernard Ghanem

In this work, we propose a new training regularizer that aims to minimize the probabilistic expected training loss of a DNN subject to a generic Gaussian input.

Data Augmentation

AdvPC: Transferable Adversarial Perturbations on 3D Point Clouds

1 code implementation ECCV 2020 Abdullah Hamdi, Sara Rojas, Ali Thabet, Bernard Ghanem

Our proposed attack increases the attack success rate by up to 40% for those transferred to unseen networks (transferability), while maintaining a high success rate on the attacked network.

Adversarial Attack Classify 3D Point Clouds

Assessing the Robustness of Visual Question Answering Models

no code implementations30 Nov 2019 Jia-Hong Huang, Modar Alfadly, Bernard Ghanem, Marcel Worring

In this work, we propose a new method that uses semantically related questions, dubbed basic questions, acting as noise to evaluate the robustness of VQA models.

Question Answering Visual Question Answering +1

Self-Supervised Learning by Cross-Modal Audio-Video Clustering

1 code implementation NeurIPS 2020 Humam Alwassel, Dhruv Mahajan, Bruno Korbar, Lorenzo Torresani, Bernard Ghanem, Du Tran

To the best of our knowledge, XDC is the first self-supervised learning method that outperforms large-scale fully-supervised pretraining for action recognition on the same architecture.

Audio Classification Deep Clustering +4

PointRGCN: Graph Convolution Networks for 3D Vehicles Detection Refinement

no code implementations27 Nov 2019 Jesus Zarzar, Silvio Giancola, Bernard Ghanem

We integrate residual GCNs in a two-stage 3D object detection pipeline, where 3D object proposals are refined using a novel graph representation.

3D Object Detection Autonomous Driving +1

G-TAD: Sub-Graph Localization for Temporal Action Detection

5 code implementations CVPR 2020 Mengmeng Xu, Chen Zhao, David S. Rojas, Ali Thabet, Bernard Ghanem

In this work, we propose a graph convolutional network (GCN) model to adaptively incorporate multi-level semantic context into video features and cast temporal action detection as a sub-graph localization problem.

Ranked #14 on Temporal Action Localization on THUMOS’14 (mAP IOU@0.5 metric)

Temporal Action Localization

DeepGCNs: Making GCNs Go as Deep as CNNs

4 code implementations15 Oct 2019 Guohao Li, Matthias Müller, Guocheng Qian, Itzel C. Delgadillo, Abdulellah Abualshour, Ali Thabet, Bernard Ghanem

This work transfers concepts such as residual/dense connections and dilated convolutions from CNNs to GCNs in order to successfully train very deep GCNs.

3D Point Cloud Classification 3D Semantic Segmentation +1

Expected Tight Bounds for Robust Deep Neural Network Training

no code implementations25 Sep 2019 Salman AlSubaihi, Adel Bibi, Modar Alfadly, Abdullah Hamdi, Bernard Ghanem

al. that bounded input intervals can be inexpensively propagated from layer to layer through deep networks.

On the Decision Boundaries of Deep Neural Networks: A Tropical Geometry Perspective

no code implementations25 Sep 2019 Motasem Alfarra, Adel Bibi, Hasan Hammoud, Mohamed Gaafar, Bernard Ghanem

We use tropical geometry, a new development in the area of algebraic geometry, to provide a characterization of the decision boundaries of a simple neural network of the form (Affine, ReLU, Affine).

Network Pruning

Finding Moments in Video Collections Using Natural Language

2 code implementations30 Jul 2019 Victor Escorcia, Mattia Soldan, Josef Sivic, Bernard Ghanem, Bryan Russell

We evaluate our approach on two recently proposed datasets for temporal localization of moments in video with natural language (DiDeMo and Charades-STA) extended to our video corpus moment retrieval setting.

Moment Retrieval Re-Ranking +2

Constrained K-means with General Pairwise and Cardinality Constraints

1 code implementation24 Jul 2019 Adel Bibi, Baoyuan Wu, Bernard Ghanem

In this paper, we enforce the above two categories into a unified clustering model starting with the integer program formulation of the standard K-means.

Expected Tight Bounds for Robust Training

2 code implementations28 May 2019 Salman Al-Subaihi, Adel Bibi, Modar Alfadly, Abdullah Hamdi, Bernard Ghanem

In this paper, we closely examine the bounds of a block of layers composed in the form of Affine-ReLU-Affine.

MAP Inference via L2-Sphere Linear Program Reformulation

1 code implementation9 May 2019 Baoyuan Wu, Li Shen, Tong Zhang, Bernard Ghanem

Thus, LS-LP is equivalent to the original MAP inference problem.

Rethinking the Pipeline of Demosaicing, Denoising and Super-Resolution

1 code implementation7 May 2019 Guocheng Qian, Yuanhao Wang, Chao Dong, Jimmy S. Ren, Wolfgang Heidrich, Bernard Ghanem, Jinjin Gu

Such a mixture problem is usually solved by a sequential solution (applying each method independently in a fixed order: DM $\to$ DN $\to$ SR), or is simply tackled by an end-to-end network without enough analysis into interactions among tasks, resulting in an undesired performance drop in the final image quality.

Demosaicking Denoising +1

Deep Layers as Stochastic Solvers

no code implementations ICLR 2019 Adel Bibi, Bernard Ghanem, Vladlen Koltun, Rene Ranftl

In particular, we show that a forward pass through a standard dropout layer followed by a linear layer and a non-linear activation is equivalent to optimizing a convex optimization objective with a single iteration of a $\tau$-nice Proximal Stochastic Gradient method.

Analytical Moment Regularizer for Gaussian Robust Networks

1 code implementation24 Apr 2019 Modar Alfadly, Adel Bibi, Bernard Ghanem

Despite the impressive performance of deep neural networks (DNNs) on numerous vision tasks, they still exhibit yet-to-understand uncouth behaviours.

Data Augmentation

Learning a Controller Fusion Network by Online Trajectory Filtering for Vision-based UAV Racing

no code implementations18 Apr 2019 Matthias Müller, Guohao Li, Vincent Casser, Neil Smith, Dominik L. Michels, Bernard Ghanem

A common approach is to learn an end-to-end policy that directly predicts controls from raw images by imitating an expert.

IAN: Combining Generative Adversarial Networks for Imaginative Face Generation

no code implementations16 Apr 2019 Abdullah Hamdi, Bernard Ghanem

Generative Adversarial Networks (GANs) have gained momentum for their ability to model image distributions.

Face Generation

BAOD: Budget-Aware Object Detection

no code implementations10 Apr 2019 Alejandro Pardo, Mengmeng Xu, Ali Thabet, Pablo Arbelaez, Bernard Ghanem

We adopt a hybrid supervised learning framework to train the object detector from both these types of annotation.

Active Learning object-detection +1

ThumbNet: One Thumbnail Image Contains All You Need for Recognition

no code implementations10 Apr 2019 Chen Zhao, Bernard Ghanem

Although deep convolutional neural networks (CNNs) have achieved great success in computer vision tasks, its real-world application is still impeded by its voracious demand of computational resources.

Towards Analyzing Semantic Robustness of Deep Neural Networks

1 code implementation9 Apr 2019 Abdullah Hamdi, Bernard Ghanem

Despite the impressive performance of Deep Neural Networks (DNNs) on various vision tasks, they still exhibit erroneous high sensitivity toward semantic primitives (e. g. object pose).

Adversarial Attack Autonomous Driving +1

DeepGCNs: Can GCNs Go as Deep as CNNs?

1 code implementation ICCV 2019 Guohao Li, Matthias Müller, Ali Thabet, Bernard Ghanem

Finally, we use these new concepts to build a very deep 56-layer GCN, and show how it significantly boosts performance (+3. 7% mIoU over state-of-the-art) in the task of point cloud semantic segmentation.

3D Semantic Segmentation Graph Classification +1

MortonNet: Self-Supervised Learning of Local Features in 3D Point Clouds

1 code implementation30 Mar 2019 Ali Thabet, Humam Alwassel, Bernard Ghanem

In fact, we show how Morton features can be used to significantly improve performance (+3% for 2 popular semantic segmentation algorithms) in the task of semantic segmentation of point clouds on the challenging and large-scale S3DIS dataset.

Self-Supervised Learning Semantic Segmentation

Efficient Bird Eye View Proposals for 3D Siamese Tracking

no code implementations25 Mar 2019 Jesus Zarzar, Silvio Giancola, Bernard Ghanem

Successively, we refine our selection of 3D object candidates by exploiting the similarity capability of a 3D Siamese network.

Object Tracking Region Proposal

SADA: Semantic Adversarial Diagnostic Attacks for Autonomous Applications

1 code implementation5 Dec 2018 Abdullah Hamdi, Matthias Müller, Bernard Ghanem

In contrast, we present a general framework for adversarial attacks on trained agents, which covers semantic perturbations to the environment of the agent performing the task as well as pixel-level attacks.

Adversarial Attack Autonomous Driving +3

SOD-MTGAN: Small Object Detection via Multi-Task Generative Adversarial Network

no code implementations ECCV 2018 Yancheng Bai, Yongqiang Zhang, Mingli Ding, Bernard Ghanem

In the MTGAN, the generator is a super-resolution network, which can up-sample small blurred images into fine-scale ones and recover detailed information for more accurate detection.

object-detection Small Object Detection +1

Face Super-resolution Guided by Facial Component Heatmaps

no code implementations ECCV 2018 Xin Yu, Basura Fernando, Bernard Ghanem, Fatih Porikli, Richard Hartley

State-of-the-art face super-resolution methods use deep convolutional neural networks to learn a mapping between low-resolution (LR) facial patterns and their corresponding high-resolution (HR) counterparts by exploring local information.

Face Hallucination Super-Resolution

The ActivityNet Large-Scale Activity Recognition Challenge 2018 Summary

no code implementations11 Aug 2018 Bernard Ghanem, Juan Carlos Niebles, Cees Snoek, Fabian Caba Heilbron, Humam Alwassel, Victor Escorcia, Ranjay Krishna, Shyamal Buch, Cuong Duc Dao

The guest tasks focused on complementary aspects of the activity recognition problem at large scale and involved three challenging and recently compiled datasets: the Kinetics-600 dataset from Google DeepMind, the AVA dataset from Berkeley and Google, and the Moments in Time dataset from MIT and IBM Research.

Activity Recognition

Diagnosing Error in Temporal Action Detectors

1 code implementation ECCV 2018 Humam Alwassel, Fabian Caba Heilbron, Victor Escorcia, Bernard Ghanem

Despite the recent progress in video understanding and the continuous rate of improvement in temporal action localization throughout the years, it is still unclear how far (or close?)

Temporal Action Localization Video Understanding

Analytic Expressions for Probabilistic Moments of PL-DNN With Gaussian Input

no code implementations CVPR 2018 Adel Bibi, Modar Alfadly, Bernard Ghanem

Moreover, we show how these expressions can be used to systematically construct targeted and non-targeted adversarial attacks.

Image Classification

Finding Tiny Faces in the Wild With Generative Adversarial Network

no code implementations CVPR 2018 Yancheng Bai, Yongqiang Zhang, Mingli Ding, Bernard Ghanem

In this paper, we proposed an algorithm to directly generate a clear high-resolution face from a blurry small one by adopting a generative adversarial network (GAN).

Face Detection

Driving Policy Transfer via Modularity and Abstraction

no code implementations25 Apr 2018 Matthias Müller, Alexey Dosovitskiy, Bernard Ghanem, Vladlen Koltun

Simulation can help end-to-end driving systems by providing a cheap, safe, and diverse training environment.

Autonomous Driving

SoccerNet: A Scalable Dataset for Action Spotting in Soccer Videos

2 code implementations12 Apr 2018 Silvio Giancola, Mohieddine Amine, Tarek Dghaily, Bernard Ghanem

A total of 6, 637 temporal annotations are automatically parsed from online match reports at a one minute resolution for three main classes of events (Goal, Yellow/Red Card, and Substitution).

Action Classification Action Detection +2

Supervised Convolutional Sparse Coding

no code implementations8 Apr 2018 Lama Affara, Bernard Ghanem, Peter Wonka

Convolutional Sparse Coding (CSC) is a well-established image representation model especially suited for image restoration tasks.

Image Reconstruction Image Restoration

Guess Where? Actor-Supervision for Spatiotemporal Action Localization

2 code implementations5 Apr 2018 Victor Escorcia, Cuong D. Dao, Mihir Jain, Bernard Ghanem, Cees Snoek

Second, we propose an actor-based attention mechanism that enables the localization of the actions from action class labels and actor proposals and is end-to-end trainable.

Action Localization Weakly Supervised Action Localization

Multi-label Learning with Missing Labels using Mixed Dependency Graphs

no code implementations31 Mar 2018 Baoyuan Wu, Fan Jia, Wei Liu, Bernard Ghanem, Siwei Lyu

This work focuses on the problem of multi-label learning with missing labels (MLML), which aims to label each test instance with multiple class labels given training instances that have an incomplete/partial set of these labels.

Image Retrieval Multi-Label Learning +1

Tagging like Humans: Diverse and Distinct Image Annotation

no code implementations CVPR 2018 Baoyuan Wu, Weidong Chen, Peng Sun, Wei Liu, Bernard Ghanem, Siwei Lyu

In D2IA, we generate a relevant and distinct tag subset, in which the tags are relevant to the image contents and semantically distinct to each other, using sequential sampling from a determinantal point process (DPP) model.

TAG

OIL: Observational Imitation Learning

no code implementations3 Mar 2018 Guohao Li, Matthias Müller, Vincent Casser, Neil Smith, Dominik L. Michels, Bernard Ghanem

Recent work has explored the problem of autonomous navigation by imitating a teacher and learning an end-to-end policy, which directly predicts controls from raw images.

Autonomous Driving Autonomous Navigation +2

Contextual Multi-Scale Region Convolutional 3D Network for Activity Detection

no code implementations28 Jan 2018 Yancheng Bai, Huijuan Xu, Kate Saenko, Bernard Ghanem

In this paper, we propose the contextual multi-scale region convolutional 3D network (CMS-RC3D) for activity detection.

Action Detection Activity Detection

A Novel Framework for Robustness Analysis of Visual QA Models

no code implementations16 Nov 2017 Jia-Hong Huang, Cuong Duc Dao, Modar Alfadly, Bernard Ghanem

In VQA, adversarial attacks can target the image and/or the proposed main question and yet there is a lack of proper analysis of the later.

Question Answering Visual Question Answering +1

ActivityNet Challenge 2017 Summary

no code implementations22 Oct 2017 Bernard Ghanem, Juan Carlos Niebles, Cees Snoek, Fabian Caba Heilbron, Humam Alwassel, Ranjay Khrisna, Victor Escorcia, Kenji Hata, Shyamal Buch

The ActivityNet Large Scale Activity Recognition Challenge 2017 Summary: results and challenge participants papers.

Activity Recognition

Constrained Convolutional Sparse Coding for Parametric Based Reconstruction of Line Drawings

no code implementations ICCV 2017 Sara Shaheen, Lama Affara, Bernard Ghanem

The process of drawing a line drawing can be approximated as the sparse spatial localization of a number of typical basic strokes, which in turn can be cast as a non-standard CSC model that considers the line drawing formation process from parametric curves.

Image Compression

High Order Tensor Formulation for Convolutional Sparse Coding

no code implementations ICCV 2017 Adel Bibi, Bernard Ghanem

Convolutional sparse coding (CSC) has gained attention for its successful role as a reconstruction and a classification tool in the computer vision and machine learning community.

Video Reconstruction

Fast Convolutional Sparse Coding in the Dual Domain

no code implementations27 Sep 2017 Lama Affara, Bernard Ghanem, Peter Wonka

Convolutional sparse coding (CSC) is an important building block of many computer vision applications ranging from image and video compression to deep learning.

Video Compression

Teaching UAVs to Race: End-to-End Regression of Agile Controls in Simulation

no code implementations19 Aug 2017 Matthias Müller, Vincent Casser, Neil Smith, Dominik L. Michels, Bernard Ghanem

Automating the navigation of unmanned aerial vehicles (UAVs) in diverse scenarios has gained much attention in recent years.

Data Augmentation Imitation Learning

Sim4CV: A Photo-Realistic Simulator for Computer Vision Applications

no code implementations19 Aug 2017 Matthias Müller, Vincent Casser, Jean Lahoud, Neil Smith, Bernard Ghanem

We present a photo-realistic training and evaluation simulator (Sim4CV) with extensive applications across various fields of computer vision.

Autonomous Driving

Learning Rotation for Kernel Correlation Filter

no code implementations11 Aug 2017 Abdullah Hamdi, Bernard Ghanem

Kernel Correlation Filters have shown a very promising scheme for visual tracking in terms of speed and accuracy on several benchmarks.

Visual Tracking

Multi-Branch Fully Convolutional Network for Face Detection

no code implementations20 Jul 2017 Yancheng Bai, Bernard Ghanem

We test our MB-FCN detector on two public face detection benchmarks, including FDDB and WIDER FACE.

Face Detection

Context-Aware Correlation Filter Tracking

no code implementations CVPR 2017 Matthias Mueller, Neil Smith, Bernard Ghanem

Correlation filter (CF) based trackers have recently gained a lot of popularity due to their impressive performance on benchmark datasets, while maintaining high frame rates.

FFTLasso: Large-Scale LASSO in the Fourier Domain

no code implementations CVPR 2017 Adel Bibi, Hani Itani, Bernard Ghanem

Since all operations in our FFTLasso method are element-wise, the subproblems are completely independent and can be trivially parallelized (e. g. on a GPU).

Dimensionality Reduction Face Recognition +2

SCC: Semantic Context Cascade for Efficient Action Detection

no code implementations CVPR 2017 Fabian Caba Heilbron, Wayner Barrios, Victor Escorcia, Bernard Ghanem

Despite the recent advances in large-scale video analysis, action detection remains as one of the most challenging unsolved problems in computer vision.

Action Detection

A Matrix Splitting Method for Composite Function Minimization

no code implementations CVPR 2017 Ganzhao Yuan, Wei-Shi Zheng, Bernard Ghanem

Incorporating a new Gaussian elimination procedure, the matrix splitting method achieves state-of-the-art performance.

Diverse Image Annotation

no code implementations CVPR 2017 Baoyuan Wu, Fan Jia, Wei Liu, Bernard Ghanem

To this end, we treat the image annotation as a subset selection problem based on the conditional determinantal point process (DPP) model, which formulates the representation and diversity jointly.

TAG

ISTA-Net: Interpretable Optimization-Inspired Deep Network for Image Compressive Sensing

1 code implementation CVPR 2018 Jian Zhang, Bernard Ghanem

With the aim of developing a fast yet accurate algorithm for compressive sensing (CS) reconstruction of natural images, we combine in this paper the merits of two existing categories of CS methods: the structure insights of traditional optimization-based methods and the speed of recent network-based ones.

Compressive Sensing

Action Search: Spotting Actions in Videos and Its Application to Temporal Action Localization

1 code implementation ECCV 2018 Humam Alwassel, Fabian Caba Heilbron, Bernard Ghanem

To address this need, we propose the new problem of action spotting in video, which we define as finding a specific action in a video while observing a small portion of that video.

Action Spotting Temporal Action Localization

VQABQ: Visual Question Answering by Basic Questions

no code implementations19 Mar 2017 Jia-Hong Huang, Modar Alfadly, Bernard Ghanem

Given a natural language question about an image, the first module takes the question as input and then outputs the basic questions of the main given question.

Question Answering Visual Question Answering +1

Fast Temporal Activity Proposals for Efficient Detection of Human Actions in Untrimmed Videos

no code implementations CVPR 2016 Fabian Caba Heilbron, Juan Carlos Niebles, Bernard Ghanem

In many large-scale video analysis scenarios, one is interested in localizing and recognizing human activities that occur in short temporal intervals within long untrimmed videos.

Action Detection Action Recognition +1

3D Part-Based Sparse Tracker With Automatic Synchronization and Registration

no code implementations CVPR 2016 Adel Bibi, Tianzhu Zhang, Bernard Ghanem

In this paper, we present a part-based sparse tracker in a particle filter framework where both the motion and appearance model are formulated in 3D.

Occlusion Handling

In Defense of Sparse Tracking: Circulant Sparse Tracker

no code implementations CVPR 2016 Tianzhu Zhang, Adel Bibi, Bernard Ghanem

Sparse representation has been introduced to visual tracking by finding the best target candidate with minimal reconstruction error within the particle filter framework.

Visual Tracking

$\ell_p$-Box ADMM: A Versatile Framework for Integer Programming

no code implementations26 Apr 2016 Baoyuan Wu, Bernard Ghanem

This paper revisits the integer programming (IP) problem, which plays a fundamental role in many computer vision and machine learning applications.

Graph Matching

Intrinsic Scene Decomposition From RGB-D images

no code implementations ICCV 2015 Mohammed Hachama, Bernard Ghanem, Peter Wonka

In this paper, we address the problem of computing an intrinsic decomposition of the colors of a surface into an albedo and a shading term.

Intrinsic Image Decomposition

What Makes an Object Memorable?

no code implementations ICCV 2015 Rachit Dubey, Joshua Peterson, Aditya Khosla, Ming-Hsuan Yang, Bernard Ghanem

We augment both the images and object segmentations from the PASCAL-S dataset with ground truth memorability scores and shed light on the various factors and properties that make an object memorable (or forgettable) to humans.

ML-MG: Multi-Label Learning With Missing Labels Using a Mixed Graph

no code implementations ICCV 2015 Baoyuan Wu, Siwei Lyu, Bernard Ghanem

This work focuses on the problem of multi-label learning with missing labels (MLML), which aims to label each test instance with multiple class labels given training instances that have an incomplete/partial set of these labels (i. e. some of their labels are missing).

Multi-Label Learning

On the Relationship Between Visual Attributes and Convolutional Networks

no code implementations CVPR 2015 Victor Escorcia, Juan Carlos Niebles, Bernard Ghanem

One of the cornerstone principles of deep models is their abstraction capacity, i. e. their ability to learn abstract concepts from `simpler' ones.

Object Recognition Zero-Shot Learning

ActivityNet: A Large-Scale Video Benchmark for Human Activity Understanding

1 code implementation CVPR 2015 Fabian Caba Heilbron, Victor Escorcia, Bernard Ghanem, Juan Carlos Niebles

In spite of many dataset efforts for human action recognition, current computer vision algorithms are still severely limited in terms of the variability and complexity of the actions that they can recognize.

Action Detection Action Recognition +2

Structural Sparse Tracking

no code implementations CVPR 2015 Tianzhu Zhang, Si Liu, Changsheng Xu, Shuicheng Yan, Bernard Ghanem, Narendra Ahuja, Ming-Hsuan Yang

Sparse representation has been applied to visual tracking by finding the best target candidate with minimal reconstruction error by use of target templates.

Visual Tracking

Robust Manhattan Frame Estimation From a Single RGB-D Image

no code implementations CVPR 2015 Bernard Ghanem, Ali Thabet, Juan Carlos Niebles, Fabian Caba Heilbron

This paper proposes a new framework for estimating the Manhattan Frame (MF) of an indoor scene from a single RGB-D image.

Representation Learning with Deep Extreme Learning Machines for Efficient Image Set Classification

no code implementations9 Mar 2015 Muhammad Uzair, Faisal Shafait, Bernard Ghanem, Ajmal Mian

Efficient and accurate joint representation of a collection of images, that belong to the same class, is a major research challenge for practical image set classification.

Classification General Classification +1

Cannot find the paper you are looking for? You can Submit a new open access paper.