
handle: 2183/31877
[Resumen]: La tecnología ligada a la Inteligencia Artificial está viviendo un enorme desarrollo en los últimos tiempos. Una de las áreas donde los avances han sido más notables es el reconocimiento de imágenes. El desarrollo de redes neuronales muy profundas permite alcanzar precisiones superiores a las de los propios humanos. Sin embargo, el entrenamiento y uso de estas grandes redes produce un importante consumo de recursos. Optimizando este uso de recursos permitimos acercar el uso de estos sistemas de reconocimiento a sistemas embebidos y móviles. El objetivo de este proyecto es mejorar la precisión de una red neuronal sencilla diseñada para el reconocimiento de imágenes, implementando algoritmos de aprendizaje por refuerzo para recortar las imágenes analizadas centrando la red neuronal de reconocimiento en la parte importante de la imagen.
[Abstract]: Technology linked to Artificial Intelligence has seen a huge development in the last few years. One of the subjects where this developments have been more notables is the field of image recognition. The development of Deep Neural Networks allows computers to achieve results better to humans. However,training and using this big networks needs a lot of computing resources. Optimizing these resources we allow to use this systems in embedded and mobile devices. The objective of this project us to improve the accuracy of a neural network designed for image recognition, by implementing a reinforcement learning algorithm to crop the analyzed images so only the important part of the image is analyzed by the classifier neural network.
Traballo fin de grao (UDC.FIC). Enxeñaría Informática. Curso 2020/2021
Reconocimiento de imagen, Artificial neural networks, Convolutional Networks, Aprendizaje profundo, Aprendizaje por refuerzo, Reinforcement learning, Redes Convolucionales, Deep learning, Image recognition, Redes Neuronales Artificiales, Artificial Neural Networks, Convolutional networks
Reconocimiento de imagen, Artificial neural networks, Convolutional Networks, Aprendizaje profundo, Aprendizaje por refuerzo, Reinforcement learning, Redes Convolucionales, Deep learning, Image recognition, Redes Neuronales Artificiales, Artificial Neural Networks, Convolutional networks
| selected citations These citations are derived from selected sources. This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically). | 0 | |
| popularity This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network. | Average | |
| influence This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically). | Average | |
| impulse This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network. | Average |
