Abstract: Sectors as fintech, legaltech or insurance process an inflow of millions of forms, invoices, id documents, claims or similar every day. Together with these, historical archives provide gigantic amounts of digitized documents containing useful information that needs to be stored in machine encoded text with a meaningful structure. This procedure, known as information extraction (IE) comprises the steps of localizing and recognizing text, identifying named entities contained in it and optionally finding relationships among its elements. In this work we explore multi-task neural models at image and graph level to solve all steps in a unified way. While doing so we find benefits and limitations of these end-to-end approaches in comparison with sequential separate methods. More specifically, we first propose a method to produce textual as well as semantic labels with a unified model from handwritten text line images. We do so with the use of a convolutional recurrent neural model trained with connectionist temporal classification to predict the textual as well as semantic information encoded in the images. Secondly, motivated by the success of this approach we investigate the unification of the localization and recognition tasks of handwritten text in full pages with an end-to-end model, observing benefits in doing so. Having two models that tackle information extraction subsequent task pairs in an end-to-end to end manner, we lastly contribute with a method to put them all together in a single neural network to solve the whole information extraction pipeline in a unified way. Doing so we observe some benefits and some limitations in the approach, suggesting that in certain cases it is beneficial to train specialized models that excel at a single challenging task of the information extraction process, as it can be the recognition of named entities or the extraction of relationships between them. For this reason we lastly study the use of the recently arrived graph neural network architectures for the semantic tasks of the information extraction process, which are recognition of named entities and relation extraction, achieving promising results on the relation extraction part.

Abstract: This paper describes the monomodal and multimodal Neural Machine Translation systems developed by LIUM and CVC for WMT17 Shared Task on Multimodal Translation. We mainly explored two multimodal architectures where either global visual features or convolutional feature maps are integrated in order to benefit from visual context. Our final systems ranked first for both En-De and En-Fr language pairs according to the automatic evaluation metrics METEOR and BLEU.

Abstract: Occupational safety is a fundamental activity in industries and revolves around the management of the necessary controls that must be present to mitigate occupational risks. These controls include verifying the use of Personal Protection Equipment (PPE). Within PPE, safety helmets are vital to reducing severe or fatal consequences caused by head injuries. This problem has been addressed recently by various research based on deep learning to detect the usage of safety helmets by the present people in the industrial field.
These works have achieved promising results for safety helmet detection using object detection methods from the YOLO family. In this work, we propose to analyze the performance of Scaled-YOLOv4, a novel model of the YOLO family that has yet to be previously studied for this problem. The performance of the Scaled-YOLOv4 is evaluated on two public databases, carefully selected among the previously proposed datasets for the occupational safety framework. We demonstrate the superiority of Scaled-YOLOv4 in terms of mAP and Fl-score concerning the previous works for both databases. Further, we summarize the currently available datasets for safety helmet detection purposes and discuss their suitability.

Abstract: Previous work from Wloka et al. (2017) presented the Selective Tuning Attentive Reference model Fixation Controller (STAR-FC), an active vision model for saccade prediction. Although the model is able to efficiently predict saccades during free-viewing, it is well known that stimulus and task instructions can strongly affect eye movement patterns (Yarbus, 1967). These factors are considered in previous Selective Tuning architectures (Tsotsos and Kruijne, 2014)(Tsotsos, Kotseruba and Wloka, 2016)(Rosenfeld, Biparva & Tsotsos 2017), proposing a way to combine bottom-up and top-down contributions to fixation and saccade programming. In particular, task priming has been shown to be crucial to the deployment of eye movements, involving interactions between brain areas related to goal-directed behavior, working and long-term memory in combination with stimulus-driven eye movement neuronal correlates. Initial theories and models of these influences include (Rao, Zelinsky, Hayhoe and Ballard, 2002)(Navalpakkam and Itti, 2005)(Huang and Pashler, 2007) and show distinct ways to process the task requirements in combination with bottom-up attention. In this study we extend the STAR-FC with novel computational definitions of Long-Term Memory, Visual Task Executive and a Task Relevance Map. With these modules we are able to use textual instructions in order to guide the model to attend to specific categories of objects and/or places in the scene. We have designed our memory model by processing a hierarchy of visual features learned from salient object detection datasets. The relationship between the executive task instructions and the memory representations has been specified using a tree of semantic similarities between the learned features and the object category labels. Results reveal that by using this model, the resulting relevance maps and predicted saccades have a higher probability to fall inside the salient regions depending on the distinct task instructions.

Abstract: We present in this paper a novel study aiming at identifying the differences in visual search patterns between physicians of diverse levels of expertise during the screening of colonoscopy videos. Physicians were clustered into two groups -experts and novices- according to the number of procedures performed, and fixations were captured by an eye-tracker device during the task of polyp search in different video sequences. These fixations were integrated into heat maps, one for each cluster. The obtained maps were validated over a ground truth consisting of a mask of the polyp, and the comparison between experts and novices was performed by using metrics such as reaction time, dwelling time and energy concentration ratio. Experimental results show a statistically significant difference between experts and novices, and the obtained maps show to be a useful tool for the characterisation of the behaviour of each group.

Abstract: Rainfall is a problem in automated traffic surveillance. Rain streaks occlude the road users and degrade the overall visibility which in turn decrease object detection performance. One way of alleviating this is by artificially removing the rain from the images. This requires knowledge of corresponding rainy and rain-free images. Such images are often produced by overlaying synthetic rain on top of rain-free images. However, this method fails to incorporate the fact that rain fall in the entire three-dimensional volume of the scene. To overcome this, we introduce training data from the SYNTHIA virtual world that models rain streaks in the entirety of a scene. We train a conditional Generative Adversarial Network for rain removal and apply it on traffic surveillance images from SYNTHIA and the AAU RainSnow datasets. To measure the applicability of the rain-removed images in a traffic surveillance context, we run the YOLOv2 object detection algorithm on the original and rain-removed frames. The results on SYNTHIA show an 8% increase in detection accuracy compared to the original rain image. Interestingly, we find that high PSNR or SSIM scores do not imply good object detection performance.

Abstract: Current visual question answering datasets do not consider the rich semantic information conveyed by text within an image. In this work, we present a new dataset, ST-VQA, that aims to highlight the importance of exploiting highlevel semantic information present in images as textual cues in the Visual Question Answering process. We use this dataset to define a series of tasks of increasing difficulty for which reading the scene text in the context provided by the visual information is necessary to reason and generate an appropriate answer. We propose a new evaluation metric for these tasks to account both for reasoning errors as well as shortcomings of the text recognition module. In addition we put forward a series of baseline methods, which provide further insight to the newly released dataset, and set the scene for further research.