Semantic segmentation of images. Using Deep Learning Techniques

      Аннотация: This book focuses on deep learning techniques in the application of semantic segmentation of images. Its objective is to clearly understand the performance of Neural Networks in different backgrounds, by training them how to learn from different images, using a deep learning-based approach. The proposed project is built around a VGG-16 convolutional neural network architecture, in order to detect the different parts and elements in an image by applying different layers, used to recognize the object in the final image. This work performs a comparative evaluation of the first pertained network against the trained one, through several repetitions and tries and different changes in its structure.Experimental results indicate how the number of times an image has been seen while training varies between the different models. These experimental results also demonstrate that segmentation methods can be very useful once trained the models.

Ключевые слова: Machine Learning, Deep Learning, Automation, Image segmentation, python, neural networks