Abstract: An effective approach to transcribe old text documents is to follow an interactive-predictive paradigm in which both, the system is guided by the human supervisor, and the supervisor is assisted by the system to complete the transcription task as efficiently as possible. In this paper, we focus on a particular system prototype called GIDOC, which can be seen as a first attempt to provide user-friendly, integrated support for interactive-predictive page layout analysis, text line detection and handwritten text transcription. More specifically, we focus on the handwriting recognition part of GIDOC, for which we propose the use of confidence measures to guide the human supervisor in locating possible system errors and deciding how to proceed. Empirical results are reported on two datasets showing that a word error rate not larger than a 10% can be achieved by only checking the 32% of words that are recognised with less confidence.

Abstract: Face detection is, in general, based on content-based detectors. Nevertheless, the face is a non-rigid object with well defined relations with respect to the human body parts. In this paper, we propose to take benefit of the context information in order to improve content-based face detections. We propose a novel framework for integrating multiple content- and context-based detectors in a discriminative way. Moreover, we develop an integrated scoring procedure that measures the ’faceness’ of each hypothesis and is used to discriminate the detection results. Our approach detects a higher rate of faces while minimizing the number of false detections, giving an average increase of more than 10% in average precision when comparing it to state-of-the art face detectors

Abstract: Conditional image synthesis is an integral part of many X2I translation systems, including image-to-image, text-to-image and audio-to-image translation systems. Training these large systems generally requires huge amounts of training data.
Therefore, we investigate knowledge distillation to transfer knowledge from a high-quality unconditioned generative model (e.g., StyleGAN) to a conditioned synthetic image generation modules in a variety of systems. To initialize the conditional and reference branch (from a unconditional GAN) we exploit the style mixing characteristics of high-quality GANs to generate an infinite supply of style-mixed triplets to perform the knowledge distillation. Extensive experimental results in a number of image generation tasks (i.e., image-to-image, semantic segmentation-to-image, text-to-image and audio-to-image) demonstrate qualitatively and quantitatively that our method successfully transfers knowledge to the synthetic image generation modules, resulting in more realistic images than previous methods as confirmed by a significant drop in the FID.

Abstract: One of the attractive characteristics of deep neural networks is their ability to transfer knowledge obtained in one domain to other related domains. As a result, high-quality networks can be trained in domains with relatively little training data. This property has been extensively studied for discriminative networks but has received significantly less attention for generative models. Given the often enormous effort required to train GANs, both computationally as well as in the dataset collection, the re-use of pretrained GANs is a desirable objective. We propose a novel knowledge transfer method for generative models based on mining the knowledge that is most beneficial to a specific target domain, either from a single or multiple pretrained GANs. This is done using a miner network that identifies which part of the generative distribution of each pretrained GAN outputs samples closest to the target domain. Mining effectively steers GAN sampling towards suitable regions of the latent space, which facilitates the posterior finetuning and avoids pathologies of other methods such as mode collapse and lack of flexibility. We perform experiments on several complex datasets using various GAN architectures (BigGAN, Progressive GAN) and show that the proposed method, called MineGAN, effectively transfers knowledge to domains with few target images, outperforming existing methods. In addition, MineGAN can successfully transfer knowledge from multiple pretrained GANs.

Abstract: This paper presents a full pipeline to classify sample sets of corn kernels. The proposed approach follows a segmentation-classification scheme. The image segmentation is performed through a well known deep learningbased approach, the Mask R-CNN architecture, while the classification is performed hrough a novel-lightweight network specially designed for this task—good corn kernel, defective corn kernel and impurity categories are considered. As a second contribution, a carefully annotated multitouching corn kernel dataset has been generated. This dataset has been used for training the segmentation and the classification modules. Quantitative evaluations have been
performed and comparisons with other approaches are provided showing improvements with the proposed pipeline.

Abstract: Oral. Paper award nominee.
Deep Neural Networks (DNNs) have revolutionized computer vision. We now have DNNs that achieve top (performance) results in many problems, including object recognition, facial expression analysis, and semantic segmentation, to name but a few. The design of the DNNs that achieve top results is, however, non-trivial and mostly done by trailand-error. That is, typically, researchers will derive many DNN architectures (i.e., topologies) and then test them on multiple datasets. However, there are no guarantees that the selected DNN will perform well in the real world. One can use a testing set to estimate the performance gap between the training and testing sets, but avoiding overfitting-to-thetesting-data is almost impossible. Using a sequestered testing dataset may address this problem, but this requires a constant update of the dataset, a very expensive venture. Here, we derive an algorithm to estimate the performance gap between training and testing that does not require any testing dataset. Specifically, we derive a number of persistent topology measures that identify when a DNN is learning to generalize to unseen samples. This allows us to compute the DNN’s testing error on unseen samples, even when we do not have access to them. We provide extensive experimental validation on multiple networks and datasets to demonstrate the feasibility of the proposed approach.

Abstract: Despite decades of research in Optical Music Recognition (OMR), the recognition of old handwritten music scores remains a challenge because of the variabilities in the handwriting styles, paper degradation, lack of standard notation, etc. Therefore, the research in OMR systems adapted to the particularities of old manuscripts is crucial to accelerate the conversion of music scores existing in archives into digital libraries, fostering the dissemination and preservation of our music heritage. In this paper we explore the adaptation of sequence-to-sequence models with attention mechanism (used in translation and handwritten text recognition) and the generation of specific synthetic data for recognizing old music scores. The experimental validation demonstrates that our approach is promising, especially when compared with long short-term memory neural networks.