Crowd Counting via Multi-view Scale Aggregation Networks

التفاصيل البيبلوغرافية
العنوان: Crowd Counting via Multi-view Scale Aggregation Networks
المؤلفون: Lingbo Liu, Qing Wang, Liang Lin, Zhilin Qiu, Nong Xiao, Guanbin Li
المصدر: ICME
بيانات النشر: IEEE, 2019.
سنة النشر: 2019
مصطلحات موضوعية: 050210 logistics & transportation, Artificial neural network, Scale (ratio), Computer science, business.industry, 05 social sciences, Feature extraction, Pattern recognition, 02 engineering and technology, Kernel (image processing), Feature (computer vision), Margin (machine learning), 0502 economics and business, 0202 electrical engineering, electronic engineering, information engineering, 020201 artificial intelligence & image processing, Artificial intelligence, business
الوصف: Crowd counting, aiming at estimating the total number of people in unconstrained crowded scenes, has increasingly received attention. But it is greatly challenged by the huge variation in people scale. In this paper, we propose a novel Multi-View Scale Aggregation Network~(MVSAN), which handle the scale variation from feature, input and criterion view comprehensively. Firstly, we design a simple but effective Multi-Scale Feature Encoder, which exploits dilated convolution layers with various dilation rates to improve the representation ability and scale diversity of features. Secondly, we feed multiple scales of input images into networks to generate high-quality density maps in a coarse-to-fine manner. Finally, we propose a Multi-Scale Structural Similarity loss to force our networks to learn the local correlation of density maps. Extensive experiments on two standard benchmarks show that the proposed method can generate high-quality crowd density map and accurate count estimation, outperforming the state-of-the-art methods with a large margin.
DOI: 10.1109/icme.2019.00259
URL الوصول: https://explore.openaire.eu/search/publication?articleId=doi_________::410670a69e2f932354571d8e03483a8b
https://doi.org/10.1109/icme.2019.00259
Rights: CLOSED
رقم الانضمام: edsair.doi...........410670a69e2f932354571d8e03483a8b
قاعدة البيانات: OpenAIRE