User Tools

Site Tools


7a.024.TUT - Real-Time Data Analytics Using Edge Computing Infrastructure with Application for Smart City

Project - Summary

This project utilities current deep learning techniques to inspect road surface and monitor road condition in a real-time. Previous works involved in such topics usually apply traditional computer vision algorithms, such as edge detection, color space segmentation or support vector machine (SVM) [1]⁠ [2]⁠, which are designed to work in centralized computing environment. Here, we expect to design a novel deep learning structure, such as Variational Auto-encoder (VAE) [3]⁠ and Generative Adversarial Networks (GANs) [4]⁠, [5]⁠ to detect road potholes and cracks as anomalies.

Project - Team

Team Member Role Email Phone Number Academic Site/IAB
Moncef Gabbouj PI +358 (400) 736613 Tampere University
Jenni Raitoharju Co-PI +358 50 447 8418 Tampere University
Lei Xu Researcher (Student) Not available Not available Tampere University
Honglei Zhang Co-PI (Student) Not available Tampere University
Juhani Ahonen Project Mentor Not available Nokia Networks
Sponsor: Nokia Networks

Project - Novelty of Approach

  • We aim to design machine learning algorithms that balance the computing complexity and data transfer cost.
  • The designed algorithms have the potential to be distributed in multiple nearby units and these units can cooperate to complete a complex task.
  • Design autonomous road condition inspection, assessment and reporting

Project - Deliverables

1 Finding references and proper datasets
2 A novel adversarial or auto-encoder structure for road anomalies detection
3 Implementing this work in a real-time
4 Publish a paper about the whole work

Project - Benefits to IAB

The final results have shown that deep generative models are well-suited for anomaly detection tasks, but the networks need to be simplified to achieve real-time goal. Hence, we plan to simplify the networks and add our novel ideas for the future publication.

Project - Presentation Video

Project - Documents

FilenameFilesizeLast modified
7a.024.tut_ip_info_sheet.docx114.5 KiB2019/08/20 11:47
7a.024.tut_final_report.pdf516.9 KiB2019/08/20 09:14
7a.024.tut_cvdi_mid-year-report_lei_honglei.pdf312.8 KiB2019/08/13 15:04
7a.024.tut_object_tracking_using_high-speed_quad_2017_fall_meeting.pptx668.3 KiB2019/08/13 15:04
7a.024.tut_confluence_project_page.pdf686.1 KiB2019/08/13 15:04
7a.024.tut_quad_chart_2018_spring_meeting.pptx101.9 KiB2019/08/13 15:04
7a.024.tut_executive_summary.pdf126.2 KiB2019/08/13 15:04
projects/year7/7a.024.tut.txt · Last modified: 2021/06/02 15:10 by sally.johnson