An Analysis on the Effect of Skip Connections in Fully Convolutional Networks for License Plate Localization

Date

2019-01-01

Author

Uzun, Engin
Akagündüz, Erdem

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

91
views

0
downloads

In this study, the effect of the skip connections, which are seen in fully convolutional networks, on object localization is analyzed. For this purpose, a local data set for plate detection is created. Experiments are carried out using this data set. Due to the small size of the image set, data augmentation method is used to overcome the danger of over-fitting. The learning rates of the first layers are frozen for analysis and fine-tuning is applied to only the last layer and deconvolution layers. The results obtained are compared with the results of other image sets. The results indicate the importance of the information provided by the skip connections on object localization.

URI

https://hdl.handle.net/11511/94013

DOI

https://doi.org/10.1109/siu.2019.8806344

Conference Name

27th Signal Processing and Communications Applications Conference (SIU)

Collections

Graduate School of Informatics, Conference / Seminar

Suggestions

OpenMETU
Core

An Experimental investigation on indoor RSSI-based localization Karakurt, Meriç Koray; Yılmaz, Ali Özgür; Department of Electrical and Electronics Engineering (2013) In this study, received signal strength indicator (RSSI) based indoor localization is investigated in sparse-anchor-deployment environments. Multipath propagation, dynamical variations in propagation model parameters and antenna patterns which are three of many potential error sources of indoor RSSI-based localization are experimentally analyzed. Possible enhancements so as to minimize the effects of mentioned problems are examined. A multichannel maximum-likelihood estimation (MLE) algorithm is proposed an...
An Analysis for the Correlation of Coverage and Spatial Resolution for Wireless Sensor Networks Tomur, Emrah; Erten, Y. Murat (2007-12-10) In this study, we investigate the interactions between coverage and spatial resolution for cluster-based wireless sensor networks (WSN). We present an approximate probabilistic analysis for the mentioned correlation and verify this analysis by simulation. Our analysis includes the k-coverage case.
A Comparison of comprehensibility between controlled Turkish and graphical representations Canpolat Şahin, Müge; Zeyrek Bozşahin, Deniz; Bozşahin, Hüseyin Cem; Department of Cognitive Sciences (2013) There are studies which show that visual representations, given in appropriate formats consistent with the context, are more informative and easy to understand as compared to textual representations (e.g., Simon & Larkin, 1980; Schnotz & Bannert, 2003; Carney & Levin, 2002). However, the factors affecting the comprehensibility of visual and textual representations should be taken into account before coming up with this conclusion. The quality of a text is closely related to the linguistic abilities of its w...
Investigation of a method to identify energy efficient organization of compression unit in parallel tree multipliers Rashid, Muhammad Saleh; Muhtaroğlu, Ali; Sustainable Environment and Energy Systems (2018-1) In this study, high-performance integer multipliers extensively used in digital signal processing are investigated in the context of energy-aware system organization. Among different functional blocks in an integer multiplier, the compression unit is primarily targeted for optimization as the largest section of the multiplier with significant energy consumption. Five different realizations of the most popular Wallace tree are investigated in this work, which consists of Dual Pass Logic (DPL) circuit impleme...
On localization and tracking using received signal strength measurements / Yılmaz, Alptekin; Yılmaz, Ali Özgür; Orguner, Umut; Department of Electrical and Electronics Engineering (2015) In this study, first, some received signal strength (RSS) based localization techniques, including maximum likelihood estimation (MLE), multidimensional scaling (MDS) and weighted least squares (WLS), are investigated and compared to each other via a simulation study within the perspective of a collaborative localization scenario. MLE using RSS measurement model, called RSS-MLE is known in the literature to be significantly biased. An important observation of this work is that the aforementioned bias can be...

Citation Formats

E. Uzun and E. Akagündüz, “An Analysis on the Effect of Skip Connections in Fully Convolutional Networks for License Plate Localization,” presented at the 27th Signal Processing and Communications Applications Conference (SIU), Sivas, Türkiye, 2019, Accessed: 00, 2021. [Online]. Available: https://hdl.handle.net/11511/94013.