Attended to WACV 2020 in Aspen, Colorado
Research stay in University of Trento
Presentation at DLBCN 2019
The Deep Learning Barcelona Symposium aims at showcasing research on deep learning carried out by scientists in/from Barcelona. There I presented my WACV article Exploring Hate Speech Detection in Multimodal Publications.
Poster and tutorial at ICDAR 2019 in Sydney
I attended to ICDAR 2019 in Sydney, were I presented my Selective Style Transfer for Text poster.
I also participated in the tutorial organized by my research group: Vision and Language: the text modality in computer vision.
[Tutorial Slides] [Poster] [Paper]
Demo at Mobile World Congress
A demo of my work on Multi-Modal Semantic Image Retrieval was shown at Barcelona Mobile World Congress in the Computer Vision Center stand. The demo is available online here.
Attended to ECCV 2018 in Munich
I attended to ECCV 2018 where I presented an oral and a poster in the Multimodal Learning and Applicattions Workshop.
Oral: “Learning to Learn from Web Data through Deep Semantic Embeddings” [PDF] [Slides] [Poster]
Poster: “Learning from #Barcelona Instagram data what Locals and Tourists post about its Neighbourhoods” [PDF] [Poster]
Poster at International Computer Vision Summer School 2018
I went to this summer school in Sicily where I presented a poster about the work done applying self-supervised learning to Social Media Barcelona data. This work was later published in ECCV MULA Workshop.
Eurecat organized a tourism forum in Barcelona, ForumTurisTIC. I presented there the possibilities of applying my research on learning from images and associated text to tourism analysis in the inspirational session.
[Slides] [Video] [BlogPost]
Poster at IAPR Summer School on Document Analysis 2017 (Jaipur, India)
Automatic discovery of textual information in images from Social Networks using FCN
I did the computer vision master offered by the UAB (Computer Vision Center), UPC, UPF, and UOC. I did the Master’s Thesis at Eurecat. It was about reading text in Social Media images. It ended in a text/non-text image classification system and in the improvement of the state of the art TextProposals algorithm, which lead to a publication. So in a pipeline to efficiently filter images with text and extract their text regions.
[PDF] [Slides] [Poster] [Code]