Vehicle classification for large-scale traffic surveillance videos using Convolutional Neural Networks

Li, Zhuo; Jiang, Liying; Zhu, Ziqi; Li, Jiafeng; Zhang, Jing; Long, Haixia

doi:10.1007/s00138-017-0846-2

Cited by 54 publications

(29 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Seven other papers are selected in response to an open call for papers. These papers cover a wide range of subtopics of intelligent urban computing with big data, including intelligent video surveillance [4,5,11], intelligent urban sensing technologies [8,10], machine vision algorithms in urban computing [2,3,9], intelligent traffic system [7], 3D vision in urban computing [1,6].…”

mentioning

confidence: 99%

See 1 more Smart Citation

Special issue on intelligent urban computing with big data

Liu

Cui

Nurminen

et al. 2017

Machine Vision and Applications

View full text Add to dashboard Cite

The rapid proliferation of urbanization has modernized many people's lives and also engendered critical issues, such as traffic congestion, energy consumption, and environmental pollution. These urbanization challenges seriously deteriorate people's life quality in big cities. Nowadays, sensing technologies and large-scale computing infrastructures have produced a variety of big data in urban spaces, which provide rich knowledge about a city to help tackle these challenges. Consequently, the intelligent urban computing, which holistically exploits the big data in big cities to improve the urban environment, human life quality, and city operation systems, has obtained massive attention in research and industrial fields. Many efforts have been dedicated to connecting unobtrusive and ubiquitous sensing technologies, advanced data management and analytics models, and novel visualization methods to structure intelligent urban computing systems for smart cities. Furthermore, the intelligent systems and applications are emerging and becoming pervasive in the field of urban planning, transportation systems, environmental conservation, energy consumption, social applications, economy, public security, and presenting representative scenarios.The goal of this special issue is to call for a coordinated effort to understand the opportunities and challenges emerging in intelligent urban computing with big data, identify key tasks, and evaluate the state-of-the-art methods, showcase innovative methodologies and ideas, introduce interesting real-world intelligent urban computing systems or applications, propose new real-world datasets and discuss future directions. We solicit original contributions in all fields of intelligent urban computing that explore the big data in big cities to help us understand the nature of urban phenomena and even predict the future of cities.With the assistance of professional referees, 11 papers out from 33 submissions are accepted after at least two rounds of rigorous reviews. Four of them were developed from conference contributions at IEEE CCIS 2016 [4,5,7,10]. Seven other papers are selected in response to an open call for papers. These papers cover a wide range of subtopics of intelligent urban computing with big data, including intelligent video surveillance [4,5,11], intelligent urban sensing technologies [8,10], machine vision algorithms in urban computing [2,3,9], intelligent traffic system [7], 3D vision in urban computing [1,6].In the first paper, Huang et al. [4] propose a novel human body segmentation framework based on shape constraint. The shape constraint is the combination of human star convexity and body parts' locations with high precision. The experimental results demonstrate that the proposed method outperforms many state-of-the-art methods on public challenging datasets. Li et al. [5] propose a new person re-identification framework. They firstly design a Siamese Inception network to 123

show abstract

mentioning

confidence: 99%

“…Zhuo et al [11] propose a vehicle classification dataset named "VehicleDataset," which contains 13,700 images extracted from real surveillance videos in various changes of resolutions, illumination, noise, angle of video cameras, and weather. The vehicles are classified into six categories, i.e., bus, car, motorcycle, minibus, truck, and van.…”

mentioning

confidence: 99%

Special issue on intelligent urban computing with big data

Liu

Cui

Nurminen

et al. 2017

Machine Vision and Applications

View full text Add to dashboard Cite

show abstract

“…Unlike previous trackers, more emphasis is put on unsupervised feature learning. A noteworthy performance improvement in visual tracking is observed with the introduction of deep neural networks (DNN) [269,270] and convolutional neural networks (CNN) [271][272][273][274][275]. DNN, especially CNN, demonstrate a strong efficiency in learning feature representations from huge annotated visual data unlike handcrafted features.…”

Section: Discussionmentioning

confidence: 99%

Vision Tracking: A Survey of the State-of-the-Art

et al. 2020

View full text Add to dashboard Cite

Vision tracking is a well-studied framework in vision computing. Developing a robust visual tracking system is challenging because of the sudden change in object motion, cluttered background, partial occlusion and camera motion. In this study, the state-of-the art visual tracking methods are reviewed and different categories are discussed. The overall visual tracking process is divided into four stages-object initialization, appearance modeling, motion estimation, and object localization. Each of these stages is briefly elaborated and related researches are discussed. A rapid growth of visual tracking algorithms is observed in last few decades. A comprehensive review is reported on different performance metrics to evaluate the efficiency of visual tracking algorithms which might help researchers to identify new avenues in this area. Various application areas of the visual tracking are also discussed at the end of the study.

show abstract

“…The feature description for animal detection used was a combination of deep learning (pretrained caffe CNN) and oriented gradient histogram features encoded with Fisher vectors. (Zhuo et al, 2017) fine-tuned their own vehicle dataset using GoogLeNet, pretrained with ILSVRC-2012 data, to obtain vehicle classification results. (Yao et al, 2017) have detected vehicles using Bayesian probability model and classified multive- (b) While most of the available research focus on using different networks and classifiers for specific applications, some researchers have particularly focused on increasing speed and accuracy while reducing false alarms for these.…”

Section: Related Workmentioning

confidence: 99%

A convolutional feature map-based deep network targeted towards traffic detection and classification

Kaur

Bhattacharya

2019

Expert Systems with Applications

View full text Add to dashboard Cite

This research mainly emphasizes on traffic detection thus essentially involving object detection and classification. The particular work discussed here is motivated from unsatisfactory attempts of re-using well known pre-trained object detection networks for domain specific data. In this course, some trivial issues leading to prominent performance drop are identified and ways to resolve them are discussed. For example, some simple yet relevant tricks regarding data collection and sampling prove to be very beneficial. Also, introducing a blur net to deal with blurred real time data is another important factor promoting performance elevation. We further study the neural network design issues for beneficial object classification and involve shared, region-independent convolutional features. Adaptive learning rates to deal with saddle points are also investigated and an average covariance matrix based pre-conditioned approach is proposed.We also introduce the use of optical flow features to accommodate orientation information. Experimental results demonstrate that this results in a steady rise in the performance rate.

show abstract

Vehicle classification for large-scale traffic surveillance videos using Convolutional Neural Networks

Cited by 54 publications

References 17 publications

Special issue on intelligent urban computing with big data

Special issue on intelligent urban computing with big data

Vision Tracking: A Survey of the State-of-the-Art

A convolutional feature map-based deep network targeted towards traffic detection and classification

Contact Info

Product

Resources

About