| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 6.43 MB | Adobe PDF |
Advisor(s)
Abstract(s)
Atualmente a automação de tarefas é uma prática cada vez mais comum em diversos sectores, uma vez que permite reduzir a necessidade de mão de obra e aumentar a eficiência de tarefas. O envelhecimento é um processo natural e complexo no desenvolvimento do ser humano. É afetado por fatores intrínsecos e extrínsecos. A compreensão deste processo é fundamental para viabilizar a deteção de idade baseado em características faciais. O presente trabalho propõe a construção de um sistema de deteção de idade com base numa imagem facial do utilizador. Este sistema contempla, numa fase inicial o pré-processamento da imagem seguido do desenvolvimento de um modelo de deteção de idade através de uma rede neuronal convolucional. O sistema foi ainda disponibilizado através de uma aplicação web. Dos vários modelos desenvolvidos com recurso às redes Xception, VGG-16 e Inception-V4 o que obteve melhor performance foi o modelo Xception. Este modelo, prevendo 4 faixas etárias, apresentou uma taxa de acerto de 88%.
Nowadays task automation is a common practice in many sectors. It allows to reduce labor and increase efficiency. Aging is a natural and complex process in the human being development. This process is affected by intrinsic and extrinsic factors. The comprehension of this process is fundamental to detect age based on facial characteristics. For this project we propose the creation of an aging detection system, based on a user's facial image. The system comprises, in an initial phase, the pre-processing of the image followed by the development of an age detection model through a convolutional neural network. The system will be available through a web application. There were many models created based on the Xception, VGG-16 and Inception-V4 networks, but the one with the best performance was an Xception model. This model, assessing 4 age groups, had and accuracy of 88%.
Nowadays task automation is a common practice in many sectors. It allows to reduce labor and increase efficiency. Aging is a natural and complex process in the human being development. This process is affected by intrinsic and extrinsic factors. The comprehension of this process is fundamental to detect age based on facial characteristics. For this project we propose the creation of an aging detection system, based on a user's facial image. The system comprises, in an initial phase, the pre-processing of the image followed by the development of an age detection model through a convolutional neural network. The system will be available through a web application. There were many models created based on the Xception, VGG-16 and Inception-V4 networks, but the one with the best performance was an Xception model. This model, assessing 4 age groups, had and accuracy of 88%.
Description
Keywords
Idade Deteção Redes Neuronais Convolucionais Web app Age Detection Convolutional neural networks