arXiv 2208.070752023-01-30cs.CV

Crowd Counting on Heavily Compressed Images with Curriculum Pre-Training

Aarhus University（奥胡斯大学）

Arian Bakhtiarnia, Qi Zhang, Alexandros Iosifidis

英文摘要：

JPEG image compression algorithm is a widely used technique for image size reduction in edge and cloud computing settings. However, applying such lossy compression on images processed by deep neural networks can lead to significant accuracy degradation. Inspired by the curriculum learning paradigm, we propose a training approach called curriculum pre-training (CPT) for crowd counting on compressed images, which alleviates the drop in accuracy resulting from lossy compression. We verify the effectiveness of our approach by extensive experiments on three crowd counting datasets, two crowd counting DNN models and various levels of compression. The proposed training method is not overly sensitive to hyper-parameters, and reduces the error, particularly for heavily compressed images, by up to 19.70%.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。

阅读原文

阅读 PDF ↗

arXiv 页面 ↗HTML 原文 ↗

本文内容

英文摘要相关论文

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。

继续探索

相关论文

Learning Error-Driven Curriculum for Crowd Counting

Academy for Engineering and Technology, Fudan University（复旦大学工程与应用技术研究院）；Shanghai Key Lab of Intelligent Information Processing, School of Computer Science, Fudan University（复旦大学计算机科学技术学院上海市智能信息处理重点实验室）

Density regression has been widely employed in crowd counting. However, the frequency imbalance of pixel values in the density map is still an obstacle to improve the performance. In this paper, we propose a novel learning strategy for learning error-driven curriculum, which uses an additional network to supervise the training of the main network. A tutoring network called TutorNet is proposed to repetitively indicate the critical errors of the main network. TutorNet generates pixel-level weights to formulate the curriculum for the main network during training, so that the main network will as…

2020-07-19cs.CV

用于人群计数的课程学习——是否值得？

Curriculum for Crowd Counting -- Is it Worthy?

Qatar Mobility Innovations Center, Qatar University（卡塔尔大学卡塔尔移动创新中心）；Qatar University（卡塔尔大学）

本文通过112组实验系统评估课程学习在人群计数中的价值，发现其能提升模型性能并加速收敛，为课程学习是否值得采用提供了实证依据。

2024-01-15cs.AIcs.CV

Crowd Flow Segmentation in Compressed Domain using CRF

Indian Institute of Science（印度科学学院）；Supercomputer Education and Research Centre（超级计算机教育与研究中心）

Crowd flow segmentation is an important step in many video surveillance tasks. In this work, we propose an algorithm for segmenting flows in H.264 compressed videos in a completely unsupervised manner. Our algorithm works on motion vectors which can be obtained by partially decoding the compressed video without extracting any additional features. Our approach is based on modelling the motion vector field as a Conditional Random Field (CRF) and obtaining oriented motion segments by finding the optimal labelling which minimises the global energy of CRF. These oriented motion segments are recursi…

2015-06-19cs.CV

高保真压缩图像的细粒度主观视觉质量评估

Fine-grained subjective visual quality assessment for high-fidelity compressed images

EPFL（洛桑联邦理工学院）；Universität Konstanz（康斯坦茨大学）；Instituto Superior Técnico（高等技术学院）；Cloudinary（Cloudinary公司）

本文针对高保真压缩图像细微视觉差异难以评估的问题，提出基于视觉刺激增强与重缩放的方法，构建了以JND为单位的高精度细粒度质量量表及众包评分数据集。

2024-10-12cs.CV

Crowd Counting via Segmentation Guided Attention Networks and Curriculum Loss

Durham University（杜伦大学）

Automatic crowd behaviour analysis is an important task for intelligent transportation systems to enable effective flow control and dynamic route planning for varying road participants. Crowd counting is one of the keys to automatic crowd behaviour analysis. Crowd counting using deep convolutional neural networks (CNN) has achieved encouraging progress in recent years. Researchers have devoted much effort to the design of variant CNN architectures and most of them are based on the pre-trained VGG16 model. Due to the insufficient expressive capacity, the backbone network of VGG16 is usually fol…

2020-08-03cs.CVcs.LGcs.MM

SR+Codec：面向视频压缩码率降低的超分辨率基准

SR+Codec: a Benchmark of Super-Resolution for Video Compression Bitrate Reduction

Lomonosov Moscow State University（罗蒙诺索夫莫斯科国立大学）；MSU Institute for Artificial Intelligence（莫斯科国立大学人工智能研究所）

该研究构建覆盖五种主流编解码器的 SR+Codec 基准，评估19个SR模型在强压缩视频上的细节恢复、伪影与主观质量，证明SR可在质量损失较小的情况下降低码率。

2024-12-04cs.CVeess.IV

更多相关论文收起更多论文

远离人群：基于地理隔离的可扩展自监督学习

Far from the Crowd: Scalable Self-Supervised Learning via Geographic Isolation

AIKO

该研究提出基于地理隔离的无标签代理指标，将其集成到MoCoV2和MAE预训练中，在CopernicusBench的三个下游任务上验证，可大幅降低预计算成本并提升下游性能。

2026-08-21cs.CVcs.LG

Point in, Box out: Beyond Counting Persons in Crowds

College of Computer Science, Sichuan University（四川大学计算机学院）；Univ Rennes, Inria, CNRS, IRISA（雷恩大学、法国国家信息与自动化研究所、法国国家科学研究中心、雷恩第一大学信息与应用数学研究所）

Modern crowd counting methods usually employ deep neural networks (DNN) to estimate crowd counts via density regression. Despite their significant improvements, the regression-based methods are incapable of providing the detection of individuals in crowds. The detection-based methods, on the other hand, have not been largely explored in recent trends of crowd counting due to the needs for expensive bounding box annotations. In this work, we instead propose a new deep detection network with only point supervision required. It can simultaneously detect the size and location of human heads and co…

2019-04-03cs.CV

面向高保真学习型图像压缩的主观视觉质量评估

Subjective Visual Quality Assessment for High-Fidelity Learning-Based Image Compression

University of Konstanz（康斯坦茨大学）；Cloudinary（Cloudinary公司）；Huawei（华为公司）；IST-IT（IST-IT机构）

本文基于 JPEG AIC-3 方法对 JPEG AI 压缩图像开展大规模主观质量评估，构建 JND 质量量表并检验客观指标与人类感知的一致性，发现多数指标在高保真范围内过于乐观，同时将 Meng-Rosenthal-Rubin 检验引入体验质量研究。

2025-04-10cs.CVeess.IV

LCDnet: A Lightweight Crowd Density Estimation Model for Real-time Video Surveillance

Qatar Mobility Innovations Center (QMIC)（卡塔尔移动创新中心）；Qatar University（卡塔尔大学）

Automatic crowd counting using density estimation has gained significant attention in computer vision research. As a result, a large number of crowd counting and density estimation models using convolution neural networks (CNN) have been published in the last few years. These models have achieved good accuracy over benchmark datasets. However, attempts to improve the accuracy often lead to higher complexity in these models. In real-time video surveillance applications using drones with limited computing resources, deep models incur intolerable higher inference delay. In this paper, we propose …

2023-02-10cs.AIcs.CV

CLIP：用更少数据更快训练

CLIP: Train Faster with Less Data

Qatar Mobility Innovations Center (QMIC)（卡塔尔移动创新中心）；Qatar University（卡塔尔大学）

本文提出 CLIP 方法，将损失感知的数据集剪枝嵌入课程学习，通过迭代剔除低重要性样本来加速模型收敛并提升泛化能力，并在人群密度估计任务中验证了其有效性。

2023-08-02cs.AIcs.CV

面向图像压缩的恰可察觉差定位研究

Localization of Just Noticeable Difference for Image Compression

imec research group at Ghent University（根特大学imec研究团队）；University of Dundee（邓迪大学）；University of Konstanz（康斯坦茨大学）

本研究针对图像压缩中逐图恰可察觉差及关键区域定位需求，提出众包实验框架，构建含300张源图的KonJND++数据集，验证了框架的有效性与可扩展性。

2023-06-13cs.CVcs.MMeess.IV

从众包中深度学习：星系形态分类的基础

Deep learning from the crowd Fundamentals of morphological galaxy classification

Instituto Nacional de Astrofísica, Óptica y Electrónica（国家天体物理、光学与电子学研究所）；Universidad de la Laguna（拉古纳大学）；Universidad Autónoma de Tlaxcala（特拉斯卡拉自治大学）；Instituto de Astrofísica de Canarias（加那利群岛天体物理研究所）

本研究利用Galaxy Zoo 1众包数据训练CNN进行星系形态分类，发现训练全网络、利用层级结构和集成可提升准确率，高一致性下准确率超99%。

2026-09-09astro-ph.GAcs.CVcs.LG

AIM 2024压缩视频质量评估挑战赛：方法与结果

AIM 2024 Challenge on Compressed Video Quality Assessment: Methods and Results

本文报告了AIM 2024压缩视频质量评估挑战赛的结果，基于459个视频和14种编解码器评估VQA方法性能，并公开数据集与基准。

2024-10-22cs.CVcs.MMeess.IV

Swift-Image：探索紧凑统一图像生成模型的性能前沿

Exploring the Performance Frontier of Compact Unified Image Generation Models

本研究提出紧凑统一图像生成模型Swift-Image，采用6B单流DiT等技术，实现领先综合性能，压缩后3B模型无明显损失，还总结了相关实用经验。

2026-08-24cs.CV

Co-Correcting: Noise-tolerant Medical Image Classification via mutual Label Correction

Beijing University of Chemical Technology（北京化工大学）；China-Japan Friendship Hospital（中日友好医院）

With the development of deep learning, medical image classification has been significantly improved. However, deep learning requires massive data with labels. While labeling the samples by human experts is expensive and time-consuming, collecting labels from crowd-sourcing suffers from the noises which may degenerate the accuracy of classifiers. Therefore, approaches that can effectively handle label noises are highly desired. Unfortunately, recent progress on handling label noise in deep learning has gone largely unnoticed by the medical image. To fill the gap, this paper proposes a noise-tol…

2021-09-11cs.AIcs.CVcs.LG