Comparative Analysis of Deep Learning Models for Vehicle Detection
Abstract
Deep Learning techniques are now widely used instead of traditional Computer Vision. There are many Deep Learning model algorithms for each use case such as Object Detection has several models, including Faster R-CNN, SSD, and YOLO v3. The performance and results of each Deep Learning model have advantages and disadvantages. Therefore, we must determine which model is suitable for the use cases and datasets that we have so that we can make the best Deep Learning model. Based on this need, this paper will make a comparative analysis of the Deep Learning model for Vehicle Detection (the spesific of Object Detection) from the models mentioned, namely, Faster R-CNN, SSD (Single Shot Detector), and YOLO v3 (You Only Look Once) to see the advantages and the disadvantages and which ones are the best. And after a comparison, it was concluded that of the three models mentioned only YOLO v3 model is able to be used as real time detection because it has low latency due to YOLO v3 only performs single convolution process so that it makes the process simpler and faster without reduce the accuracy.
Downloads
References
Ammar, Adel., Koubaa1, Anis., Ahmed, Mohanned., Saad, Abdulrahman., 2020. Aerial Images Processing for Car Detection using Convolutional Neural Networks:Comparison between Faster R-CNN and YoloV3.
Chen, L., Ye, Feiyue., Ruan, Y., Fan, Honghui., Chen, Qimei., 2018. An Algorithm for Highway Vehicle Detection Based on Convolutional Neural Network
Deep Learning. [wikipedia] (Updated May 5, 2020) Available at : https://en.wikipedia.org/wiki/Deep_learning
Dutta, Suvajit., CS Manideep, Bonthala., Rai, Shalva., V, Vijayarajan.: A Comparative Study of Deep Learning Models for Medical Image Classification. (2017)
Hui, Jonathan., SSD object detection: Single Shot MultiBox Detector for real-time processing. [Medium] (Updated Mar 14, 2018) Available at: https://medium.com/@jonathan_hui/ssd-object-detection-single-shot-multibox-detector-for-real-time-processing-9bd8deac0e06
Kathuria, Ayoosh., 2018. What's new in YOLO v3? [Towards Data Science] (Updated Apr 23, 2018) Available at: https://towardsdatascience.com/yolo-v3-object-detection-53fb7d3bfe6b
Kazakov, I., 2017. Vehicle Detection and Tracking. [Towards Data Science] (Updated May 14, 2017) Available at: https://towardsdatascience.com/vehicle-detection-and-tracking-44b851d70508
Kumar, Prince., Garg, Vaibhav., Somvanshi, Pavan., Pathanjali.: A Comparative Study of Object Detection Algorithms in A Scene. (2019)
Liu, Wei., Anguelov, Dragomir., Erhan, Dumitru., Szegedy, Christian., Reed, Scott., Fu1, Cheng-Yang., C. Berg, Alexander.: SSD: Single Shot MultiBox Detector. (2016)
Redmon, J., Divvala, S., Girshick, R., Farhadi, A.: You only look once: Unified, real-time object detection. In: CVPR. (2016)
Redmon, J., Farhadi, Ali.: YOLOv3: An Incremental Improvement. (2018)
Ren, Shaoqing., He, Kaiming., Girshick, Ross., Sun, Jian.: Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks. (2016)
Shatnawi, Ali., Al-Bdour, Ghadeer., Al-Qurran, Raffi., Al-Ayyoub, Mahmoud.: A Comparative Study of Open Source Deep Learning Frameworks. (2018)
Xiao, Y., 2019. Vehicle Detection in Deep Learning. MSc. Virginia: Virginia Polythecnic Institute and State University
[15] Xu, Joyce., 2017. Deep Learning for Object Detection: A Comprehensive Review. [Towards Data Science] (Updated Sep 12, 2017) Available at: https://towardsdatascience.com/deep-learning-for-object-detection-a-comprehensive-review-73930816d8d9
Xu, Yinghan., Faster R-CNN (object detection) implemented by Keras for custom data from Google’s Open Images Dataset V4. [Towards Data Science] (Updated Nov 20, 2018) Available at : https://towardsdatascience.com/faster-r-cnn-object-detection-implemented-by-keras-for-custom-data-from-googles-open-images-125f62b9141a
Copyright (c) 2022 Rendi Nurcahyo
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under Creative Commons Attribution 4.0 International License that allows others to share the work with an acknowledgment of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgment of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (Refer to The Effect of Open Access).
