Search Results for author: Mohammad Sabokrou

Found 42 papers, 13 papers with code

Class-Adaptive Sampling Policy for Efficient Continual Learning

1 code implementation27 Nov 2023 Hossein Rezaei, Mohammad Sabokrou

Continual learning (CL) aims to acquire new knowledge while preserving information from previous experiences without forgetting.

Continual Learning

Explainability of Vision Transformers: A Comprehensive Review and New Perspectives

no code implementations12 Nov 2023 Rojina Kashefi, Leili Barekatain, Mohammad Sabokrou, Fatemeh Aghaeipoor

Transformers have had a significant impact on natural language processing and have recently demonstrated their potential in computer vision.

Decision Making

Mitigating Bias: Enhancing Image Classification by Improving Model Explanations

no code implementations4 Jul 2023 Raha Ahmadi, Mohammad Javad Rajabi, Mohammad Khalooie, Mohammad Sabokrou

In this paper, we propose a novel approach to address this issue and improve the learning of main concepts by image classifiers.

Classification Image Classification

Adversarial Backdoor Attack by Naturalistic Data Poisoning on Trajectory Prediction in Autonomous Driving

no code implementations27 Jun 2023 Mozhgan PourKeshavarz, Mohammad Sabokrou, Amir Rasouli

In autonomous driving, behavior prediction is fundamental for safe motion planning, hence the security and robustness of prediction models against adversarial attacks are of paramount importance.

Autonomous Driving Backdoor Attack +3

Global-Local Processing in Convolutional Neural Networks

1 code implementation14 Jun 2023 Zahra Rezvani, Soroor Shekarizeh, Mohammad Sabokrou

We devise a simple module called Global Advantage Stream (GAS) to learn and capture the holistic features of input samples (i. e., the global features).

Revealing Model Biases: Assessing Deep Neural Networks via Recovered Sample Analysis

no code implementations10 Jun 2023 Mohammad Mahdi Mehmanchi, Mahbod Nouri, Mohammad Sabokrou

While a generalization test is one way to evaluate a trained model's performance, it can be costly and may not cover all scenarios to ensure that the model has learned the primary concepts.

Test

A Unified Concept-Based System for Local, Global, and Misclassification Explanations

no code implementations6 Jun 2023 Fatemeh Aghaeipoor, Dorsa Asgarian, Mohammad Sabokrou

Explainability of Deep Neural Networks (DNNs) has been garnering increasing attention in recent years.

Quantifying Overfitting: Evaluating Neural Network Performance through Analysis of Null Space

no code implementations30 May 2023 Hossein Rezaei, Mohammad Sabokrou

To achieve this, we analyze the null space in the last layer of neural networks, which enables us to quantify overfitting without access to training data or knowledge of the accuracy of those data.

Towards Realistic Out-of-Distribution Detection: A Novel Evaluation Framework for Improving Generalization in OOD Detection

1 code implementation20 Nov 2022 Vahid Reza Khazaie, Anthony Wong, Mohammad Sabokrou

This paper presents a novel evaluation framework for Out-of-Distribution (OOD) detection that aims to assess the performance of machine learning models in more realistic settings.

Out-of-Distribution Detection Out of Distribution (OOD) Detection

MobileDenseNet: A new approach to object detection on mobile devices

1 code implementation22 Jul 2022 Mohammad Hajizadeh, Mohammad Sabokrou, Adel Rahmani

We also developed a light neck FCPNLite for mobile devices that will aid with the detection of small objects.

object-detection Object Detection

Fake It Till You Make It: Towards Accurate Near-Distribution Novelty Detection

1 code implementation28 May 2022 Hossein Mirzaei, Mohammadreza Salehi, Sajjad Shahabi, Efstratios Gavves, Cees G. M. Snoek, Mohammad Sabokrou, Mohammad Hossein Rohban

Effectiveness of our method for both the near-distribution and standard novelty detection is assessed through extensive experiments on datasets in diverse applications such as medical images, object classification, and quality control.

Ranked #2 on Anomaly Detection on One-class CIFAR-10 (using extra training data)

Anomaly Detection Novelty Detection

Deep-Disaster: Unsupervised Disaster Detection and Localization Using Visual Data

1 code implementation31 Jan 2022 Soroor Shekarizadeh, Razieh Rastgoo, Saif Al-Kuwari, Mohammad Sabokrou

In this paper, inspired by the success of Knowledge Distillation (KD) methods, we propose an unsupervised deep neural network to detect and localize damages in social media images.

Humanitarian Knowledge Distillation

All You Need In Sign Language Production

no code implementations5 Jan 2022 Razieh Rastgoo, Kourosh Kiani, Sergio Escalera, Vassilis Athitsos, Mohammad Sabokrou

To make an easy and mutual communication between the hearing-impaired and the hearing communities, building a robust system capable of translating the spoken language into sign language and vice versa is fundamental.

Cultural Vocal Bursts Intensity Prediction Sign Language Production +2

A Unified Survey on Anomaly, Novelty, Open-Set, and Out-of-Distribution Detection: Solutions and Future Challenges

1 code implementation26 Oct 2021 Mohammadreza Salehi, Hossein Mirzaei, Dan Hendrycks, Yixuan Li, Mohammad Hossein Rohban, Mohammad Sabokrou

To date, several research domains tackle the problem of detecting unfamiliar samples, including anomaly detection, novelty detection, one-class learning, open set recognition, and out-of-distribution detection.

Anomaly Detection Novelty Detection +2

Looking Back on Learned Experiences For Class/task Incremental Learning

no code implementations ICLR 2022 Mozhgan PourKeshavarzi, Guoying Zhao, Mohammad Sabokrou

In this paper, we shed light on an on-call transfer set to provide past experiences whenever a new task arises in the data stream.

Incremental Learning

ClaRe: Practical Class Incremental Learning By Remembering Previous Class Representations

no code implementations29 Mar 2021 Bahram Mohammadi, Mohammad Sabokrou

Learning new knowledge in the absence of data instances from previous classes or even imbalance samples of both old and new classes makes CIL an ongoing challenging problem.

Class Incremental Learning Incremental Learning

Sign Language Production: A Review

1 code implementation29 Mar 2021 Razieh Rastgoo, Kourosh Kiani, Sergio Escalera, Mohammad Sabokrou

Sign Language is the dominant yet non-primary form of communication language used in the deaf and hearing-impaired community.

Sign Language Production Sign Language Recognition

ZS-IL: Looking Back on Learned ExperiencesFor Zero-Shot Incremental Learning

no code implementations22 Mar 2021 Mozhgan PourKeshavarz, Mohammad Sabokrou

In this paper, we shed light on an on-call transfer set to provide past experiences whenever a new class arises in the data stream.

Incremental Learning

Ano-Graph: Learning Normal Scene Contextual Graphs to Detect Video Anomalies

no code implementations18 Mar 2021 Masoud Pourreza, Mohammadreza Salehi, Mohammad Sabokrou

Video anomaly detection has proved to be a challenging task owing to its unsupervised training procedure and high spatio-temporal complexity existing in real-world scenarios.

Anomaly Detection Graph Learning +2

Image/Video Deep Anomaly Detection: A Survey

no code implementations2 Mar 2021 Bahram Mohammadi, Mahmood Fathy, Mohammad Sabokrou

AD strongly correlates with the important computer vision and image processing tasks such as image/video anomaly, irregularity and sudden event detection.

Anomaly Detection Event Detection

Robustification of Segmentation Models Against Adversarial Perturbations In Medical Imaging

no code implementations23 Sep 2020 Hanwool Park, Amirhossein Bayat, Mohammad Sabokrou, Jan S. Kirschke, Bjoern H. Menze

This paper presents a novel yet efficient defense framework for segmentation models against adversarial attacks in medical imaging.

Segmentation

G2D: Generate to Detect Anomaly

no code implementations20 Jun 2020 Masoud Pourreza, Bahram Mohammadi, Mostafa Khaki, Samir Bouindour, Hichem Snoussi, Mohammad Sabokrou

Previous researches solve this problem as a One-Class Classification (OCC) task where they train a reference model on all of the available samples.

Binary Classification One-Class Classification

Deep-HR: Fast Heart Rate Estimation from Face Video Under Realistic Conditions

no code implementations12 Feb 2020 Mohammad Sabokrou, Masoud Pourreza, Xiaobai Li, Mahmood Fathy, Guoying Zhao

In this paper, we propose a simple yet efficient approach to benefit the advantages of the Deep Neural Network (DNN) by simplifying HR estimation from a complex task to learning from very correlated representation to HR.

Heart rate estimation regression

Code-Bridged Classifier (CBC): A Low or Negative Overhead Defense for Making a CNN Classifier Robust Against Adversarial Attacks

no code implementations16 Jan 2020 Farnaz Behnia, Ali Mirzaeian, Mohammad Sabokrou, Sai Manoj, Tinoosh Mohsenin, Khaled N. Khasawneh, Liang Zhao, Houman Homayoun, Avesta Sasan

In this paper, we propose Code-Bridged Classifier (CBC), a framework for making a Convolutional Neural Network (CNNs) robust against adversarial attacks without increasing or even by decreasing the overall models' computational complexity.

Denoising Image Classification

AutoIDS: Auto-encoder Based Method for Intrusion Detection System

no code implementations8 Nov 2019 Mohammed Gharib, Bahram Mohammadi, Shadi Hejareh Dastgerdi, Mohammad Sabokrou

These detectors are two encoder-decoder neural networks that are forced to provide a compressed and a sparse representation from the normal flows.

Intrusion Detection Test

Self-Supervised Representation Learning via Neighborhood-Relational Encoding

no code implementations ICCV 2019 Mohammad Sabokrou, Mohammad Khalooei, Ehsan Adeli

Conventional unsupervised learning methods only focused on training deep networks to understand the primitive characteristics of the visual data, mainly to be able to reconstruct the data from a latent space.

Anomaly Detection Representation Learning +1

AVD: Adversarial Video Distillation

no code implementations12 Jul 2019 Mohammad Tavakolian, Mohammad Sabokrou, Abdenour Hadid

The key idea is to represent videos by compressing them in the form of realistic images, which can be used in a variety of video-based scene analysis applications.

Activity Recognition General Classification +1

End-to-End Adversarial Learning for Intrusion Detection in Computer Networks

no code implementations25 Apr 2019 Bahram Mohammadi, Mohammad Sabokrou

Inspired by the successes of Generative Adversarial Networks (GANs) for training deep models in semi-unsupervised setting, we have proposed an end-to-end deep architecture for IDS.

Intrusion Detection One-Class Classification

Online Signature Verification using Deep Representation: A new Descriptor

no code implementations24 Jun 2018 Mohammad Hajizadeh Saffar, Mohsen Fayyaz, Mohammad Sabokrou, Mahmood Fathy

To deal with these difficulties and modeling the signatures efficiently, we propose a method that a one-class classifier per each user is built on discriminative features.

One-class classifier

Semantic Video Segmentation: A Review on Recent Approaches

no code implementations16 Jun 2018 Mohammad Hajizadeh Saffar, Mohsen Fayyaz, Mohammad Sabokrou, Mahmood Fathy

This paper gives an overview on semantic segmentation consists of an explanation of this field, it's status and relation with other vision fundamental tasks, different datasets and common evaluation parameters that have been used by researchers.

Segmentation Semantic Segmentation +2

AVID: Adversarial Visual Irregularity Detection

2 code implementations24 May 2018 Mohammad Sabokrou, Masoud Pourreza, Mohsen Fayyaz, Rahim Entezari, Mahmood Fathy, Jürgen Gall, Ehsan Adeli

Real-time detection of irregularities in visual data is very invaluable and useful in many prospective applications including surveillance, patient monitoring systems, etc.

Anomaly Detection

Adversarially Learned One-Class Classifier for Novelty Detection

5 code implementations CVPR 2018 Mohammad Sabokrou, Mohammad Khalooei, Mahmood Fathy, Ehsan Adeli

Our architecture is composed of two deep networks, each of which trained by competing with each other while collaborating to understand the underlying concept in the target class, and then classify the testing samples.

Novelty Detection One-Class Classification +2

Towards Principled Design of Deep Convolutional Networks: Introducing SimpNet

1 code implementation17 Feb 2018 Seyyed Hossein Hasanpour, Mohammad Rouhani, Mohsen Fayyaz, Mohammad Sabokrou, Ehsan Adeli

SimpNet outperforms the deeper and more complex architectures such as VGGNet, ResNet, WideResidualNet \etc, on several well-known benchmarks, while having 2 to 25 times fewer number of parameters and operations.

Image Classification

Lets keep it simple, Using simple architectures to outperform deeper and more complex architectures

9 code implementations22 Aug 2016 Seyyed Hossein Hasanpour, Mohammad Rouhani, Mohsen Fayyaz, Mohammad Sabokrou

Our simple 13-layer architecture outperforms most of the deeper and complex architectures to date such as VGGNet, ResNet, and GoogleNet on several well-known benchmarks while having 2 to 25 times fewer number of parameters and operations.

Image Classification

STFCN: Spatio-Temporal FCN for Semantic Video Segmentation

1 code implementation21 Aug 2016 Mohsen Fayyaz, Mohammad Hajizadeh Saffar, Mohammad Sabokrou, Mahmood Fathy, Reinhard klette, Fay Huang

Current work on convolutional neural networks(CNNs) has shown that CNNs provide advanced spatial features supporting a very good performance of solutions for both image and video analysis, especially for the semantic segmentation task.

Segmentation Semantic Segmentation +2

Real-Time Anomaly Detection and Localization in Crowded Scenes

no code implementations21 Nov 2015 Mohammad Sabokrou, Mahmood Fathy, Mojtaba Hosseini, Reinhard klette

In this paper, we propose a method for real-time anomaly detection and localization in crowded scenes.

Anomaly Detection

IDSA: Intelligent Distributed Sensor Activation Algorithm For Target Tracking With Wireless Sensor Network

no code implementations30 May 2015 Mohammad Sabokrou, Mahmood Fathy, Mojtaba Hoseini

One important application of the Wireless Sensor Network(WSN) is target tracking, the aim of this application is converging to an event or object in an area.

Feature Representation for Online Signature Verification

no code implementations29 May 2015 Mohsen Fayyaz, Mohammad Hajizadeh_Saffar, Mohammad Sabokrou, Mahmood Fathy

Biometrics systems have been used in a wide range of applications and have improved people authentication.

feature selection

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