Awesome Open Source
Awesome Open Source

Traffic Sign Detection

This is the code for the paper

Evaluation of deep neural networks for traffic sign detection systems
Álvaro Arcos-García, Juan Antonio Álvarez-García, Luis M. Soria-Morillo

The paper addresses the problem of traffic sign detection analysing the state-of-the-art of several object-detection systems (Faster R-CNN, R-FCN, SSD, and YOLO V2) combined with various feature extractors (Resnet V1 50, Resnet V1 101, Inception V2, Inception Resnet V2, Mobilenet V1, and Darknet-19). We aim to explore the properties of these object-detection models which are modified and specifically adapted to the traffic sign detection problem domain by means of transfer learning. In particular, various publicly available object-detection models that were pre-trained on the Microsoft COCO dataset are fine-tuned on the German Traffic Sign Detection Benchmark dataset. The evaluation and comparison of these models include key metrics, such as the mean average precision (mAP), memory allocation, running time, number of floating point operations, number of parameters of the model, and the effect of traffic sign image sizes. Detection example

We provide:

If you find this code useful in your research, please cite:

"Evaluation of deep neural networks for traffic sign detection systems."
Álvaro Arcos-García, Juan A. Álvarez-García, Luis M. Soria-Morillo. Neurocomputing 316 (2018) 332-344.

[link][bibtex]

Requirements

This project is implemented in Tensorflow and it is based on two repositories: Tensorflow Object Detection API and darkflow.

Pretrained models

You can download TFRecords of the GTSDB dataset from Google Drive. Unzip it to the gtsdb_data folder. You can download pre-trained models from Google Drive. Unzip them to the models folder.

Results

Overview

Our findings show that Faster R-CNN Inception Resnet V2 obtains the best mAP, while R-FCN Resnet 101 strikes the best trade-off between accuracy and execution time. YOLO V2 and SSD Mobilenet merit a special mention, in that the former achieves competitive accuracy results and is the second fastest detector, while the latter, is the fastest and the lightest model in terms of memory consumption, making it an optimal choice for deployment in mobile and embedded devices.

model mAP parameters flops memory_mb total_exec_millis accelerator_exec_millis cpu_exec_millis
Faster R-CNN Resnet 50 91.52 43337242 533575386662 5256.454615 104.0363553 75.93395395 28.10240132
Faster R-CNN Resnet 101 95.08 62381593 625779295782 6134.705805 123.2729175 90.33714433 32.9357732
Faster R-CNN Inception V2 90.62 12891249 120621363525 2175.206857 58.53338971 38.76813971 19.76525
Faster R-CNN Inception Resnet V2 95.77 59412281 1837544257834 18250.446008 442.2206796 366.1586796 76062
R-FCN Resnet 101 95.15 64594585 269898731281 3509.75153 85.45207971 52.40321739 33.04886232
SSD Mobilenet 61.64 5572809 2300721483 94.696119 15.14525 4.021267857 11.12398214
SSD Inception V2 66.10 13474849 7594247747 284.512918 23.74428378 9.393405405 14.35087838
YOLO V2 78.83 50588958 62780021160 1318.108256 21.4810122 18.13923171 3.341780488

Results plot

Image size comparison

model small medium large
Faster R-CNN Resnet 50 53.57 86.95 86.72
Faster R-CNN Resnet 101 70.89 94.17 88.87
Faster R-CNN Inception V2 56.72 81.02 88.53
Faster R-CNN Inception Resnet V2 68.60 86.62 82.10
R-FCN Resnet 101 60.37 82.03 79.56
SSD Mobilenet 22.13 55.32 82.06
SSD Inception V2 26.85 64.71 78.76
YOLO V2 42.93 78.99 75.67

mAP vs image size

Running on new images

We provide a Jupyter Notebook with instructions to run our pre-trained models on new images.

Training

Config files needed to train our models are included within each zip file that you can download from the pre-trained models section. Minimal changes were done to the base source code of tensorflow and darkflow so you should be able to reproduce our results using the last versions of the aforementioned repositories. For instance, to train a SDD Mobilenet model you can run the following command lines:

cd ~/tensorflow/models/research
python3 object_detection/train.py --logtostderr --pipeline_config_path models/ssd_mobilenet_v1/ssd_mobilenet_v1_gtsdb3.config --train_dir models/ssd_mobilenet_v1/train/

Evaluation

We provide a Jupyter Notebook with instructions to evaluate the results obtained by our pre-trained models.

Acknowledgements

The source code of this project is mainly based on Tensorflow Object Detection API and darkflow.

Alternatives To Traffic Sign Detection
Select To Compare


Alternative Project Comparisons
Related Awesome Lists
Top Programming Languages

Get A Weekly Email With Trending Projects For These Topics
No Spam. Unsubscribe easily at any time.
Jupyter Notebook (178,660
Deep Learning (39,275
Tensorflow (22,886
Artificial Intelligence (20,093
Convolutional Neural Networks (12,853
Sign (7,416
Object Detection (4,250
Deep Neural Networks (4,053
Resnet (3,233
Ssd (1,973
Mobilenet (1,161
Research Paper (517
Faster Rcnn (304
Yolov2 (139
Darkflow (27
Traffic Sign Detection (21
Gtsdb (4
Rfcn Resnet (4