Convolutional neural network for dog breed recognition system
The object of research is models of convolutional neural networks and datasets for training convolutional neural networks. The purpose of the study is to increase the efficiency of recognition of dog breeds using convolutional neural networks when implementing them in mobile applications. To achiev...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Igor Sikorsky Kyiv Polytechnic Institute
2024-10-01
|
| Series: | Adaptivni Sistemi Avtomatičnogo Upravlinnâ |
| Subjects: | |
| Online Access: | https://asac.kpi.ua/article/view/313022 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The object of research is models of convolutional neural networks and datasets for training convolutional neural networks. The purpose of the study is to increase the efficiency of recognition of dog breeds using convolutional neural networks when implementing them in mobile applications. To achieve the goal, a neural network model based on the modified MobileNetV3-Large model is proposed, which is trained on the revised "70 Dog Breeds-Image Data Set" using data augmentation methods. The proposed model shows a recognition accuracy of 96% on 70 different dog breeds. The resulting model is integrated into the dog breed recognition system, which is based on a mobile platform. The implemented program creates functional opportunities for interaction with the received model and identifies the breed of the dog in real time.
Ref. 30, pic. 5
|
|---|---|
| ISSN: | 1560-8956 2522-9575 |