检索结果-内蒙古大学图书馆

MARK MY WORDS: DANGERS OF WATERMARKED IMAGES IN IMAGENET

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Bykov, Kirill Müller, Klaus-Robert Höhne, Marina M.-C. Technische Universität Berlin Machine Learning Group Berlin10587 Germany Understandable Machine Intelligence Lab ATB Potsdam14469 Germany BIFOLD Berlin Institute for the Foundations of Learning and Data Berlin10587 Germany Korea University Department of Artificial Intelligence Seoul136-713 Korea Republic of Max Planck Institute for Informatics Saarbrücken66123 Germany Machine Learning Group UiT the Arctic University of Norway Tromsø9037 Norway Department of Computer Science University of Potsdam Potsdam14476 Germany

The utilization of pre-trained networks, especially those trained on ImageNet, has become a common practice in computer Vision. However, prior research has indicated that a significant number of images in the ImageNet dataset contain watermarks, making pre-trained networks susceptible to learning artifacts such as watermark patterns within their latent spaces. In this paper, we aim to assess the extent to which popular pre-trained architectures display such behavior and to determine which classes are most affected. Additionally, we examine the impact of watermarks on the extracted features. Contrary to the popular belief that the Chinese logographic watermarks impact the "carton" class only, our analysis reveals that a variety of ImageNet classes, such as "monitor", "broom", "apron" and "safe" rely on spurious correlations. Finally, we propose a simple approach to mitigate this issue in fine-tuned networks by ignoring the encodings from the feature-extractor layer of ImageNet pre-trained networks that are most susceptible to watermark imprints. © 2023, CC BY.

关键词： Watermarking

Assessment of Clinical Metadata on the Accuracy of Retinal Fundus Image labels in Diabetic Retinopathy in Uganda: Case-Crossover Study Using the Multimodal Database of Retinal Images in Africa

学校读者我要写书评

暂无评论

JMIR Formative Research 2024年 8卷 e59914页

作者： Arunga, Simon Morley, Katharine Elise Kwaga, Teddy Morley, Michael Gerard Nakayama, Luis Filipe Mwavu, Rogers Kaggwa, Fred Ssempiira, Julius Celi, Leo Anthony Haberer, Jessica E. Obua, Celestino Department of Ophthalmology Mbarara University of Science and Technology Mbarara Uganda Massachusetts General Hospital Center for Global Health Department of Medicine Harvard Medical School Boston MA United States Harvard Ophthalmology AI Lab Massachusetts Eye and Ear Infirmary Harvard Medical School Boston MA United States Ophthalmology Department Sao Paulo Federal University Sao Paulo Brazil Laboratory for Computational Physiology Massachusetts Institute of Technology Cambridge MA United States Faculty of Computing and Informatics Department of Information Technology Mbarara University of Science and Technology Mbarara Uganda Faculty of Computing and Informatics Department of Computer Science Mbarara University of Science and Technology Mbarara Uganda School of Public Health Makerere University Kampala Uganda Division of Pulmonary Critical Care and Sleep Medicine Beth Israel Deaconess Medical Center Harvard Medical School Boston MA United States Department of Biostatistics Harvard T.H. Chan School of Public Health Boston MA United States Department of Pharmacology Mbarara University of Science and Technology Mbarara Uganda

Background: labeling color fundus photos (CFP) is an important step in the development of artificial intelligence screening algorithms for the detection of diabetic retinopathy (DR). Most studies use the International Classification of Diabetic Retinopathy (ICDR) to assign labels to CFP, plus the presence or absence of macular edema (ME). Images can be grouped as referrable or nonreferrable according to these classifications. There is little guidance in the literature about how to collect and use metadata as a part of the CFP labeling process. Objective: This study aimed to improve the quality of the Multimodal Database of Retinal Images in Africa (MoDRIA) by determining whether the availability of metadata during the image labeling process influences the accuracy, sensitivity, and specificity of image labels. MoDRIA was developed as one of the inaugural research projects of the Mbarara University Data science Research Hub, part of the Data science for Health Discovery and Innovation in Africa (DS-I Africa) initiative. Methods: This is a crossover assessment with 2 groups and 2 phases. Each group had 10 randomly assigned labelers who provided an ICDR score and the presence or absence of ME for each of the 50 CFP in a test image with and without metadata including blood pressure, visual acuity, glucose, and medical history. Specificity and sensitivity of referable retinopathy were based on ICDR scores, and ME was calculated using a 2-sided t test. Comparison of sensitivity and specificity for ICDR scores and ME with and without metadata for each participant was calculated using the Wilcoxon signed rank test. Statistical significance was set at P<.05. Results: The sensitivity for identifying referrable DR with metadata was 92.8% (95% CI 87.6-98.0) compared with 93.3% (95% CI 87.6-98.9) without metadata, and the specificity was 84.9% (95% CI 75.1-94.6) with metadata compared with 88.2% (95% CI 79.5-96.8) without metadata. The sensitivity for identifying the presence of

关键词： Africa African AI algorithms artificial intelligence assessment bias clinical images diabetes diabetic diabetic retinopathy image labeling metadata multimodal database retinal images screening screening algorithms sensitivity treatment

Deep hypergraph U-Net for brain graph embedding and classification

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Lostar, Mert Rekik, Islem BASIRA lab Faculty of Computer and Informatics Istanbul Technical University Istanbul Turkey School of Science and Engineering Computing University of Dundee United Kingdom

Background. Network neuroscience examines the brain as a complex system represented by a network (or connectome), providing deeper insights into the brain morphology and function, allowing the identification of atypical brain connectivity alterations, which can be used as diagnostic markers of neurological disorders. -Existing Methods. Graph embedding methods which map data samples (e.g., brain networks) into a low dimensional space have been widely used to explore the relationship between samples for classification or prediction tasks. However, the majority of these works are based on modeling the pair-wise relationships between samples, failing to capture their higher-order relationships. -New Method. In this paper, inspired by the nascent field of geometric deep learning, we propose Hypergraph U-Net (HUNet), a novel data embedding framework leveraging the hypergraph structure to learn low-dimensional embeddings of data samples while capturing their high-order relationships. Specifically, we generalize the U-Net architecture, naturally operating on graphs, to hypergraphs by improving local feature aggregation and preserving the high-order relationships present in the data. -Results. We tested our method on small-scale and large-scale heterogeneous brain connectomic datasets including morphological and functional brain networks of autistic and demented patients, respectively. -Conclusion. Our HUNet outperformed state-of-the art geometric graph and hypergraph data embedding techniques with a gain of 4-14% in classification accuracy, demonstrating both scalability and generalizability. Copyright © 2020, The Authors. All rights reserved.

关键词： computer aided diagnosis

Machine Learning Methods for Brain Network Classification: Application to Autism Diagnosis using Cortical Morphological Networks

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Bilgen, Ismail Guvercin, Goktug Rekik, Islem BASIRA lab Faculty of Computer and Informatics Istanbul Technical University Istanbul Turkey School of Science and Engineering Computing University of Dundee United Kingdom

-Background. Autism spectrum disorder (ASD) affects the brain connectivity at different levels. Nonetheless, non-invasively distinguishing such effects using magnetic resonance imaging (MRI) remains very challenging to machine learning diagnostic frameworks due to ASD heterogeneity. So far, existing network neuroscience works mainly focused on functional (derived from functional MRI) and structural (derived from diffusion MRI) brain connectivity, which might not capture relational morphological changes between brain regions. Indeed, machine learning (ML) studies for ASD diagnosis using morphological brain networks derived from conventional T1-weighted MRI are very scarce. -New Method. To fill this gap, we leverage crowdsourcing by organizing a Kaggle competition to build a pool of machine learning pipelines for neurological disorder diagnosis with application to ASD diagnosis using cortical morphological networks derived from T1-weighted MRI. -Results. During the competition, participants were provided with a training dataset and only allowed to check their performance on a public test data. The final evaluation was performed on both public and hidden test datasets based on accuracy, sensitivity, and specificity metrics. Teams were ranked using each performance metric separately and the final ranking was determined based on the mean of all rankings. The first-ranked team achieved 70% accuracy, 72.5% sensitivity, and 67.5% specificity, while the second-ranked team achieved 63.8%, 62.5%, 65% respectively. -Conclusion. Leveraging participants to design ML diagnostic methods within a competitive machine learning setting has allowed the exploration and benchmarking of wide spectrum of ML methods for ASD diagnosis using cortical morphological networks. Copyright © 2020, The Authors. All rights reserved.

关键词： Python

SoccerNet 2024 Challenges Results

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Cioppa, Anthony Giancola, Silvio Somers, Vladimir Joos, Victor Magera, Floriane Held, Jan Ghasemzadeh, Seyed Abolfazl Zhou, Xin Seweryn, Karolina Kowalczyk, Mateusz Mróz, Zuzanna Lukasik, Szymon Halon, Michal Mkhallati, Hassan Deliège, Adrien Hinojosa, Carlos Sanchez, Karen Mansourian, Amir M. Miralles, Pierre Barnich, Olivier De Vleeschouwer, Christophe Alahi, Alexandre Ghanem, Bernard Van Droogenbroeck, Marc Gorski, Adam Clapés, Albert Boiarov, Andrei Afanasiev, Anton Xarles, Artur Scott, Atom Lim, ByoungKwon Yeung, Calvin Gonzalez, Cristian Rüfenacht, Dominic Pacilio, Enzo Deuser, Fabian Altawijri, Faisal Sami Cachón, Francisco Kim, HanKyul Wang, Haobo Choe, Hyeonmin Kim, Hyunwoo J. Kim, Il-Min Kang, Jae-Mo Tursunboev, Jamshid Yang, Jian Hong, Jihwan Lee, Jimin Zhang, Jing Lee, Junseok Zhang, Kexin Habel, Konrad Jiao, Licheng Li, Linyi Gutiérrez-Pérez, Marc Ortega, Marcelo Li, Menglong Lopatto, Milosz Kasatkin, Nikita Nemtsev, Nikolay Oswald, Norbert Udin, Oleg Kononov, Pavel Geng, Pei Alotaibi, Saad Ghazai Kim, Sehyung Ulasen, Sergei Escalera, Sergio Zhang, Shanshan Yang, Shuyuan Moon, Sunghwan Moeslund, Thomas B. Shandyba, Vasyl Golovkin, Vladimir Dai, Wei Chung, WonTaek Liu, Xinyu Zhu, Yongqiang Kim, Youngseo Li, Yuan Yang, Yuting Xiao, Yuxuan Cheng, Zehua Li, Zhihao Belgium King Abdullah University of Science and Technology Saudi Arabia Sportradar Switzerland UCLouvain Belgium EPFL Switzerland EVS Broadcast Equipment Belgium Baidu Research United States NASK National Research Institute Poland Belgium Sharif University of Technology Iran Footovision France Mobius Labs Institut de Robòtica i Informàtica Industrial CSIC-UPC Spain TAHAKOM Saudi Arabia AIBrain Eidos.ai United Kingdom School of Computer Science and Engineering Nanjing University of Science and Technology China Constructor.tech Universitat de Barcelona Spain Computer Vision Center Spain Aalborg University Denmark Kyungpook National University Korea Republic of Intelligent Perception and Image Understanding Lab Xidian University China Germany Korea University Korea Republic of Warsaw University of Technology Poland University of Oxford United Kingdom China Science IntelliCloud Technology Co. Ltd China Robo Space Graduate School of Informatics Nagoya University Japan Tongji University China AGH University of Krakow Poland

The SoccerNet 2024 challenges represent the fourth annual video understanding challenges organized by the SoccerNet team. These challenges aim to advance research across multiple themes in football, including broadcast video understanding, field understanding, and player understanding. This year, the challenges encompass four vision-based tasks. (1) Ball Action Spotting, focusing on precisely localizing when and which soccer actions related to the ball occur, (2) Dense Video Captioning, focusing on describing the broadcast with natural language and anchored timestamps, (3) Multi-View Foul Recognition, a novel task focusing on analyzing multiple viewpoints of a potential foul incident to classify whether a foul occurred and assess its severity, (4) Game State Reconstruction, another novel task focusing on reconstructing the game state from broadcast videos onto a 2D top-view map of the field. Detailed information about the tasks, challenges, and leaderboards can be found at https://***, with baselines and development kits available at https://***/SoccerNet. © 2024, CC BY.

关键词： HTTP

Enhanced Outsourced and Secure Inference for Tall Sparse Decision Trees

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Quijano, Andrew Halkidis, Spyros T. Gallagher, Kevin Akkaya, Kemal Samaras, Nikolaos Dept. of Computer Science and Engineering New York University New YorkNY United States Dept. of Applied Informatics University of Macedonia Thessaloniki Greece NOVA LINCS NOVA School of Science and Technology Lisbon Portugal Advanced Wireless and Security Lab Florida International University MiamiFL United States

A decision tree is an easy-to-understand tool that has been widely used for classification tasks. On the one hand, due to privacy concerns, there has been an urgent need to create privacy-preserving classifiers that conceal the user’s input from the classifier. On the other hand, with the rise of cloud computing, data owners are keen to reduce risk by outsourcing their model, but want security guarantees that third parties cannot steal their decision tree model. To address these issues, Joye and Salehi introduced a theoretical protocol that efficiently evaluates decision trees while maintaining privacy by leveraging their comparison protocol that is resistant to timing attacks. However, their approach was not only inefficient but also prone to side-channel attacks. Therefore, in this paper, we propose a new decision tree inference protocol in which the model is shared and evaluated among multiple entities. We partition our decision tree model by each level to be stored in a new entity we refer to as a "level-site." Utilizing this approach, we were able to gain improved average run time for classifier evaluation for a non-complete tree, while also having strong mitigations against side-channel attacks. Copyright © 2025, The Authors. All rights reserved.

关键词： Side channel attack

FooBaR: Fault Fooling Backdoor Attack on Neural Network Training

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Breier, Jakub Hou, Xiaolu Ochoa, Martín Solano, Jesus Silicon Austria Labs TU-Graz SAL DES Lab Graz University of Technology Graz Austria Faculty of Informatics and Information Technologies Slovak University of Technology Slovakia Department of Computer Science ETH Zurich Zurich Switzerland Appgate Inc. Bogotá Colombia

Neural network implementations are known to be vulnerable to physical attack vectors such as fault injection attacks. As of now, these attacks were only utilized during the inference phase with the intention to cause a misclassification. In this work, we explore a novel attack paradigm by injecting faults during the training phase of a neural network in a way that the resulting network can be attacked during deployment without the necessity of further faulting. In particular, we discuss attacks against ReLU activation functions that make it possible to generate a family of malicious inputs, which are called fooling inputs, to be used at inference time to induce controlled misclassifications. Such malicious inputs are obtained by mathematically solving a system of linear equations that would cause a particular behaviour on the attacked activation functions, similar to the one induced in training through faulting. We call such attacks fooling backdoors as the fault attacks at training phase inject backdoors into the network that allow an attacker to produce fooling inputs. We evaluate our approach against multi-layer perceptron networks and convolutional networks on a popular image classification task obtaining high attack success rates (from 60% to 100%) and high classification confidence when as little as 25 neurons are attacked, while preserving high accuracy on the originally intended classification task. © 2021, CC BY.

关键词： Chemical activation

Advanced Leaf Classification using Multi-Layer Perceptron for Smart Crop Management

学校读者我要写书评

暂无评论

IEEE Access 2025年

作者： Mumtaz, Sara Alshehri, Mohammed AlQahtani, Yahya Alshahrani, Abdulmonem Alabdullah, Bayan Alhasson, Haifa F. Liu, Hui Air University Department of Computer Science Islamabad44000 Pakistan King Khalid University Department of Computer Science Abha61421 Saudi Arabia King Khalid University Department of Informatics and Computer Systems Abha61421 Saudi Arabia Princess Nourah bint Abdulrahman University College of Computer and Information Sciences Department of Information Systems P.O. Box 84428 Riyadh11671 Saudi Arabia Qassim University College of Computer Department of Information Technology Buraydah52571 Saudi Arabia University of Bremen Cognitive Systems Lab Bremen28359 Germany

Accurate leaf classification plays a vital role in ecological studies and precision agriculture, yet the diversity in leaf morphology and environmental influences presents a significant challenge. To improve accuracy, this study suggests a revolutionary methodology that combines sophisticated preprocessing, segmentation, feature extraction, optimization, and classification algorithms. To improve image quality, preprocessing is first done with bilateral filtering and Contrast Limited Adaptive Histogram Equalization (CLAHE). Then, to ensure accurate border identification, Conditional Random Fields (CRF) are used for segmentation. To extract features, complex texture and shape details are captured using Local Ternary Patterns (LTP), KAZE, and Histogram of Oriented Gradients (HOG). The optimization method, Stochastic Gradient Descent (SGD), refines the feature space for better categorization. Finally, a Multi-Layer Perceptron (MLP) classifier is used, and it produces remarkable results on three different datasets: accuracy of 92.63%, 88.3%, and 94.5%, respectively. This strong architecture shows great promise for disease monitoring and plant identification, providing a flexible and scalable solution for a range of agricultural applications. © 2013 IEEE.

关键词： Feature Selection

NTIRE 2023 Quality Assessment of Video Enhancement Challenge

学校读者我要写书评

暂无评论

NTIRE 2023 Quality Assessment of Video Enhancement Challenge

2023 IEEE/CVF Conference on computer Vision and Pattern Recognition Workshops, CVPRW 2023

作者： Liu, Xiaohong Min, Xiongkuo Sun, Wei Zhang, Yulun Zhang, Kai Timofte, Radu Zhai, Guangtao Gao, Yixuan Cao, Yuqin Kou, Tengchuan Dong, Yunlong Jia, Ziheng Li, Yilin Wu, Wei Hu, Shuming Deng, Sibin Xiao, Pengxiang Chen, Ying Li, Kai Zhao, Kai Yuan, Kun Sun, Ming Cong, Heng Wang, Hao Fu, Lingzhi Zhang, Yusheng Zhang, Rongyu Shi, Hang Xu, Qihang Xiao, Longan Ma, Zhiliang Agarla, Mirko Celona, Luigi Rota, Claudio Schettini, Raimondo Huang, Zhiwei Li, Ya'nan Wang, Xiaotao Lei, Lei Liu, Hongye Hong, Wei Chuang, Ironhead Lin, Allen Guan, Drake Chen, Iris Lou, Kae Huang, Willy Tasi, Yachun Kao, Yvonne Fan, Haotian Kong, Fangyuan Zhou, Shiqi Liu, Hao Lai, Yu Chen, Shanshan Wang, Wenqi Wu, Haoning Chen, Chaofeng Zhu, Chunzheng Guo, Zekun Zhao, Shiling Yin, Haibing Wang, Hongkui Meftah, Hanene Brachemi Fezza, Sid Ahmed Hamidouche, Wassim Déforges, Olivier Shi, Tengfei Mansouri, Azadeh Motamednia, Hossein Bakhtiari, Amir Hossein Aznaveh, Ahmad Mahmoudi Shanghai Jiao Tong University China Eth Zürich Switzerland University of Würzburg Germany Department of Tao Technology Alibaba Group China Kuaishou Technology China Interactive Entertainment Group of Netease Inc Guangzhou China Transsion China Hefei University of Technology China Department of Informatics Systems and Communication University of Milano - Bicocca Italy Xiaomi Inc. China China Ji Liang University China FreeTech Advanced Research Center KKCompany Taiwan ByteDance China Mgtv China Fuzhou University China Shopee Information Technology Co. Ltd. S-Lab Nanyang Technological University Singapore Hunan University China Hangzhou Dianzi University China Insa Rennes Cnrs Ietr - Umr 6164 Rennes France National Higher School of Telecommunications and Ict Oran Algeria Buaa Department of Electrical and Computer Engineering Faculty of Engineering Kharazmi University Tehran Iran High Performance Computing Laboratory School of Computer Science Institute for Research in Fundamental Sciences Tehran Iran Cyberspace Research Institute Shahid Beheshti University Tehran Iran

ISBN: (纸本)9798350302493

This paper reports on the NTIRE 2023 Quality Assessment of Video Enhancement Challenge, which will be held in conjunction with the New Trends in Image Restoration and Enhancement Workshop (NTIRE) at CVPR 2023. This challenge is to address a major challenge in the field of video processing, namely, video quality assessment (VQA) for enhanced videos. The challenge uses the VQA Dataset for Perceptual Video Enhancement (VDPVE), which has a total of 1211 enhanced videos, including 600 videos with color, brightness, and contrast enhancements, 310 videos with deblurring, and 301 deshaked videos. The challenge has a total of 167 registered participants. 61 participating teams submitted their prediction results during the development phase, with a total of 3168 submissions. A total of 176 submissions were submitted by 37 participating teams during the final testing phase. Finally, 19 participating teams submitted their models and fact sheets, and detailed the methods they used. Some methods have achieved better results than baseline methods, and the winning methods have demonstrated superior prediction performance. © 2023 IEEE.

关键词： Image enhancement