Classification of rare traffic signs
The paper studies the possibility of using neural networks for the classification of objects that are few or absent at all in the training set. The task is illustrated by the example of classification of rare traffic signs. We consider neural networks trained using a contrastive loss function and it...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Samara National Research University
2020-04-01
|
| Series: | Компьютерная оптика |
| Subjects: | |
| Online Access: | http://computeroptics.smr.ru/KO/PDF/KO44-2/440213.pdf |
| _version_ | 1818965301842149376 |
|---|---|
| author | Boris Faizov Vladislav Shakhuro Vadim Sanzharov Anton Konushin |
| author_facet | Boris Faizov Vladislav Shakhuro Vadim Sanzharov Anton Konushin |
| author_sort | Boris Faizov |
| collection | DOAJ |
| description | The paper studies the possibility of using neural networks for the classification of objects that are few or absent at all in the training set. The task is illustrated by the example of classification of rare traffic signs. We consider neural networks trained using a contrastive loss function and its modifications, also we use different methods for generating synthetic samples for classification problems. As a basic method, the indexing of classes using neural network features is used. A comparison is made of classifiers trained with three different types of synthetic samples and their mixtures with real data. We propose a method of classification of rare traffic signs using a neural network discriminator of rare and frequent signs. The experimental evaluation shows that the proposed method allows rare traffic signs to be classified without significant loss of frequent sign classification quality. |
| first_indexed | 2024-12-20T13:14:50Z |
| format | Article |
| id | doaj.art-3d253ff848fa4e1db7ad7a739f110a0b |
| institution | Directory Open Access Journal |
| issn | 0134-2452 2412-6179 |
| language | English |
| last_indexed | 2024-12-20T13:14:50Z |
| publishDate | 2020-04-01 |
| publisher | Samara National Research University |
| record_format | Article |
| series | Компьютерная оптика |
| spelling | doaj.art-3d253ff848fa4e1db7ad7a739f110a0b2022-12-21T19:39:35ZengSamara National Research UniversityКомпьютерная оптика0134-24522412-61792020-04-0144223624310.18287/2412-6179-CO-601Classification of rare traffic signsBoris Faizov0Vladislav Shakhuro1Vadim Sanzharov2Anton Konushin3Lomonosov Moscow State University, Moscow, RussiaLomonosov Moscow State University, Moscow, RussiaLomonosov Moscow State University, Moscow, Russia; Gubkin RSU of Oil and GasLomonosov Moscow State University, Moscow, Russia; NRU Higher School of Economics, Moscow, RussiaThe paper studies the possibility of using neural networks for the classification of objects that are few or absent at all in the training set. The task is illustrated by the example of classification of rare traffic signs. We consider neural networks trained using a contrastive loss function and its modifications, also we use different methods for generating synthetic samples for classification problems. As a basic method, the indexing of classes using neural network features is used. A comparison is made of classifiers trained with three different types of synthetic samples and their mixtures with real data. We propose a method of classification of rare traffic signs using a neural network discriminator of rare and frequent signs. The experimental evaluation shows that the proposed method allows rare traffic signs to be classified without significant loss of frequent sign classification quality.http://computeroptics.smr.ru/KO/PDF/KO44-2/440213.pdftraffic sign classificationsynthetic training samplesneural networksimage recognitionimage transformsneural network compositions |
| spellingShingle | Boris Faizov Vladislav Shakhuro Vadim Sanzharov Anton Konushin Classification of rare traffic signs Компьютерная оптика traffic sign classification synthetic training samples neural networks image recognition image transforms neural network compositions |
| title | Classification of rare traffic signs |
| title_full | Classification of rare traffic signs |
| title_fullStr | Classification of rare traffic signs |
| title_full_unstemmed | Classification of rare traffic signs |
| title_short | Classification of rare traffic signs |
| title_sort | classification of rare traffic signs |
| topic | traffic sign classification synthetic training samples neural networks image recognition image transforms neural network compositions |
| url | http://computeroptics.smr.ru/KO/PDF/KO44-2/440213.pdf |
| work_keys_str_mv | AT borisfaizov classificationofraretrafficsigns AT vladislavshakhuro classificationofraretrafficsigns AT vadimsanzharov classificationofraretrafficsigns AT antonkonushin classificationofraretrafficsigns |