TRAFFIC SIGN DETECTION USING
CONVOLUTIONAL NEURAL
NETWORK
19BCI0086 – CHIRUMAMILLA VISHAL
19BCB0079 – KOLLA RAVITEJA
19BCI0119 – SANAPALA TARUN
ABSTRACT
✔
Traffic-sign recognition (TSR) is a technology by which a vehicle is abletorecognize the traffic signs put
on the road e.g. "speed limit" or "children" or "turnahead". This is part of thefeatures collectively called
ADAS. The technologyis being developed by a variety of automotive suppliers. It uses imageprocessing
techniques to detect the traffic signs.
✔
The detection methods canbegenerally divided into color based, shape based and learning
basedmethods. We propose an approach for traffic sign detection based on Convolutional
NeuralNetworks (CNN). We first transform the original image into the gray scaleimageby using support
vector machines, then use convolutional neural networkswithfixed and learnable layers for detection
and recognition.
✔
The fixed layer can reduce the amount of interest areas to detect, andcropthe boundariesvery close to
the borders of traffic signs. The learnable layerscanincrease the accuracy of detection significantly
2
PROBLEM STATEMENT
✔ Traffic sign recognition (TSR) represents an important feature of advanced driver assistance systems,
contributing to the safety of the drivers, pedestrians and vehicles as well. Developing TSR systems
requires the use of computer.
✔ Vision techniques, which could be considered fundamental in the field of pattern recognition in general.
Despite all the previous works and research that has been achieved, traffic sign detection and
recognition still remains a very challenging problem, precisely if we want to provide a real time
processing solution.
3
LITERATURE SURVERY
SL TITLE OF AUTHOR YEAR OF DATA METHO PERFOM DRAWBACKS
. THE S PUBLICATI USED DLOGY ANCE
N PAPER ON METRICS
O.
1 A study on Y. Aoyagi, 22nd Traffic sign Genetic Among 24 Image pattern recognition
traffic sign T.Asakura International from a algorithm patterns has been chiefly
recognitio Conference video s and other than researched only for an
n in scene on industrial image neural the speed individual object. However
image electronics networks sign, it it is an advanced direction
using control and was to recognize the object
genetic instrumentati recognize which becomes a target
algorithms on, ieee, d that only from a scene image with
and neural aug 1996 one teh development of the
networks pattern visual system of the robot.
was not a
sign
4
LITERATURE SURVERY
SL TITLE OF AUTHOR YEAR OF DATA METHO PERFOM DRAWBACKS
. THE S PUBLICATI USED DLOGY ANCE
N PAPER ON METRICS
O.
2 Road sign Sh hsu, cl Image and 30 Matching Triangular Edges are tested at
detection huang vision triangualr pursuit road signs different levels of
and computing, road signs method 94% resolution by using so
recognitio 2001- and 10 circular called a hierarchical
n using elsevier circular road signs structure code. It is
matching road signs 91% assumed that closed edge
pursuit contours are availabe at
method one of these levels of
resolution and failures
happen when the outline
of the traffic sign merges
with the background
5
LITERATURE SURVERY
SL TITLE OF AUTHOR YEAR OF DATA METHO PERFOM DRAWBACKS
. THE S PUBLICATI USED DLOGY ANCE
N PAPER ON METRICS
O.
3 Traffic Yingying 2001 Swedish Fully Averge Color-based methods,
sign zhu, traffic signs convoluti precision shaped-based methods
detection chengqua dataset onal of 98.67% and sliding window based
and n zhang, (STSD) network methods.
recognitio duoyou guided
n using zhou, propsals
fully xinggang
convolutio wang,
n ai xiang bai,
network wenyuliu,
guided elsvie
propsals
6
LITERATURE SURVERY
SL TITLE OF AUTHOR YEAR OF DATA METHO PERFOM DRAWBACKS
. THE S PUBLICATI USED DLOGY ANCE
N PAPER ON METRICS
O.
4 Generaliz Jf khan Journal of Ladotd Neural Images Proprietary algorithms use
ed traffic sma computing in roadway networks with sign specific color filters and
sign bhuiyan civil video log in them the features of specific
detection engineering, image sets. 83.67% shapes to distinguish a
model for 2009 (37, 640 images specific type of traffic sign.
developin video log with no But they can detect only
g a sign images) sign in stop signs.
inventory them =
80.13%
7
LITERATURE SURVEY
✔ Aoyagi, Y., & Asakura, T. (1996, August). A study on traffic sign recognition in scene image using
genetic algorithms and neural networks. In Proceedings of the 1996 IEEE IECON. 22nd International
Conference on Industrial Electronics, Control, and Instrumentation (Vol. 3, pp. 1838-1843). IEEE.
✔ Hsu, S. H., & Huang, C. L. (2001). Road sign detection and recognition using matching pursuit method.
Image and Vision Computing, 19(3), 119-129.
✔ Zhu, Y., Zhang, C., Zhou, D., Wang, X., Bai, X., & Liu, W. (2016). Traffic sign detection and recognition
using fully convolutional network guided proposals. Neurocomputing, 214, 758-766.
✔ Tsai, Y., Kim, P., & Wang, Z. (2009). Generalized traffic sign detection model for developing a sign
inventory. Journal of Computing in Civil Engineering, 23(5), 266-276.
✔ Khan, J. F., Bhuiyan, S. M., & Adhami, R. R. (2010). Image segmentation and shape analysis for road-
sign detection. IEEE Transactions on Intelligent Transportation Systems, 12(1), 83-96.
