Effect of various dimension convolutional layer filters on traffic sign classification accuracy

Sichkar, Valentyn N.; Kolyubin, Sergey A.

doi:10.17586/2226-1494-2019-19-3-546-552

Cited by 5 publications

(13 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As previously introduced, in this work will be made a review of [7], so we implement the same architecture of CNN, but we plan our experiments to enable a more thorough statistical analysis of the outcomes.…”

Section: Methodsmentioning

confidence: 99%

“…The dataset used for training and evaluating the CNN performance in this work is [35] which is the same employed in [7]. Broadly speaking, it is a pre-processed derivation of the GTSRB [8], with insertion of artificially generated data to balance the number of available elements of all classes.…”

Section: A Experiments Designmentioning

confidence: 99%

“…In the following experiments, the dataset was divided between training, validation and test data, being respectively 86,989, 4,410 and 12,630, as in the reference article. However, differently from [7], each network configuration was trained from the scratch for 30 times.…”

Section: A Experiments Designmentioning

confidence: 99%

“…One exception is [7], that analyses the outcomes of distinct kernel sizes in a CNN model in the context of traffic sign recognition. However, little or no discussion is made about the character intrinsically stochastic of the learning process and the randomness of the neural network parameters initialization.…”

Section: Introductionmentioning

confidence: 99%

“…This work proposed a review of Sichkar and Kolyubin paper [7] in the direction of finding the best kernel size in a traffic signs recognition application using a CNN architecture. It aims at validating the results presented by the original paper, while proposing a more in-depth statistical analysis underlain by a broader set of experimental results.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

A Thorough Evaluation of Kernel Order in CNN Based Traffic Signs Recognition

Oliveira¹,

Mota²,

Vidal³

2020

Anais Do XVI Workshop De Visão Computacional (WVC 2020)

View full text Add to dashboard Cite

Convolutional Neural Network is an important deep learning architecture for computer vision. Alongside with its variations, it brought image analysis applications to a new performance level. However, despite its undoubted quality, the evaluation of the performance presented in the literature is mostly restricted to accuracy measurements. So, considering the stochastic characteristics of neural networks training and the impact of the architectures configuration, research is still needed to affirm if such architectures reached the optimal configuration for their focused problems. Statistical significance is a powerful tool for a more accurate experimental evaluation of stochastic processes. This paper is dedicated to perform a thorough evaluation of kernel order influence over convolutional neural networks in the context of traffic signs recognition. Experiments for distinct kernels sizes were performed using the most well accepted database, the socalled German Traffic Sign Recognition Benchmark.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: A Experiments Designmentioning

confidence: 99%

Section: A Experiments Designmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

A Thorough Evaluation of Kernel Order in CNN Based Traffic Signs Recognition

Oliveira¹,

Mota²,

Vidal³

2020

Anais Do XVI Workshop De Visão Computacional (WVC 2020)

View full text Add to dashboard Cite

show abstract

Using Proximity Graph Cut for Fast and Robust Instance‐Based Classification in Large Datasets

Protasov

Khan

2021

Complexity

View full text Add to dashboard Cite

K-nearest neighbours (kNN) is a very popular instance-based classifier due to its simplicity and good empirical performance. However, large-scale datasets are a big problem for building fast and compact neighbourhood-based classifiers. This work presents the design and implementation of a classification algorithm with index data structures, which would allow us to build fast and scalable solutions for large multidimensional datasets. We propose a novel approach that uses navigable small-world (NSW) proximity graph representation of large-scale datasets. Our approach shows 2–4 times classification speedup for both average and 99th percentile time with asymptotically close classification accuracy compared to the 1-NN method. We observe two orders of magnitude better classification time in cases when method uses swap memory. We show that NSW graph used in our method outperforms other proximity graphs in classification accuracy. Our results suggest that the algorithm can be used in large-scale applications for fast and robust classification, especially when the search index is already constructed for the data.

show abstract

Real time detection and classification of traffic signs based on YOLO version 3 algorithm

Sichkar¹,

Kolyubin²

2020

Naučno-teh. vestn. inf. tehnol. meh. opt.

Self Cite

View full text Add to dashboard Cite

The issue of effective detection and classification of various traffic signs is studied. The two-stage method is proposed for creation of holistic model with end-to-end solution. The first stage includes implementation of effective localization of traffic signs by YOLO version 3 algorithm (You Only Look Once). At the first stage the traffic signs are grouped into four categories according to their shapes. At the second stage, an accurate classification of the located traffic signs is performed into one of the forty-three predefined categories. The second stage is based on another model with one convolutional neural layer. The model for detection of traffic signs was trained on German Traffic Sign Detection Benchmark (GTSDB) with 630 and 111 RGB images for training and validation, respectively. Сlassification model was trained on German Traffic Sign Recognition Benchmark (GTSRB) with 66000 RGB images on pure "numpy" library with 19 × 19 dimension of convolutional layer filters and reached 0.868 accuracy on testing dataset. The experimental results illustrated that the training of the first model deep network with only four categories for location of traffic signs produced high mAP (mean Average Precision) accuracy reaching 97.22 %. Additional convolutional layer of the second model applied for final classification creates efficient entire system. Experiments on processing video files demonstrated frames per second (FMS) between thirty-six and sixty-one that makes the system feasible for real time applications. The frames per second depended on the number of traffic signs to be detected and classified in every single frame in the range from six to one.

show abstract

Effect of various dimension convolutional layer filters on traffic sign classification accuracy

Cited by 5 publications

References 5 publications

A Thorough Evaluation of Kernel Order in CNN Based Traffic Signs Recognition

A Thorough Evaluation of Kernel Order in CNN Based Traffic Signs Recognition

Using Proximity Graph Cut for Fast and Robust Instance‐Based Classification in Large Datasets

Real time detection and classification of traffic signs based on YOLO version 3 algorithm

Contact Info

Product

Resources

About