Both human face recognition and generation by machines are currently an active area of computer vision, drawing curiosity of researchers, capable of performing amazing image analysis, and producing applications in multiple domains. In this paper, we propose a new approach for face attributes classification (FAC) taking advantage from both binary classification and data augmentation. With binary classification we can reach high prediction scores, while augmented data prevent overfitting and overcome the lack of data for sketched photos. Our approach, named Augmented binary multilabel CNN (ABM-CNN), consists of three steps: i) splitting data; ii) transformed-it to sketch (simplification process); iii) train separately each attribute with two convolutional neural networks; the whole process includes two networks: the first (resp. the second) one is to predict attributes on real images (resp. sketches) as inputs. Through experimentation, we figure out that some attributes give high prediction rates with sketches rather than with real images. On the other hand, we build a new face dataset, more consistent and complete, by generating images using Style-GAN model, to which we apply our method for extracting face attributes. As results, our proposal demonstrates more performances compared to those of related works.
The advancements in artificial intelligence research, particularly in computer vision, have led to the development of previously unimaginable applications, such as generating new contents based on text description. In our work we focused on the text-to-image synthesis applications (TIS) field, to transform descriptive sentences into a real image. To tackle this issue, we use unsupervised deep learning networks that can generate high quality images from text descriptions, provided by eyewitnesses to assist law enforcement in their investigations, for the purpose of generating probable human faces. We analyzed a number of existing approaches and chose the best one. Deep fusion generative adversarial networks (DF-GAN) is the network that performs better than its peers, at multiple levels, like the generated image quality or the respect of the giving descriptive text. Our model is trained on the CelebA dataset and text descriptions (generated by our algorithm using existing attributes in the dataset). The obtained results from our implementation show that the learned generative model makes excellent quantitative and visual performances, the model is capable of generating realistic and diverse samples for human faces and create a complete portrait with respect of given text description.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.