Return to site

DCC UC team wins first place in international medical imaging competition.

The IALab / HAivis UC team, made up of Hans Lobel and Denis Parra, academics from the Department of Computer Science, together with undergraduate and graduate students from the Department, obtained first place in the international competition Medical ImageCLEF, category Image Captioning - Caption Prediction, organized by ImageCLEF, in addition to fourth place in the Concept Detection category and seventh place in the Tuberculosis CT analysis and Visual Question Answering categories.

ImageCLEF, part of the Conference and Labs of the Evaluation Forum (CLEF), is a space to provide support in the evaluation of independent methods for the automatic annotation of images with concepts, in addition to multimodal information retrieval methods based on factors textual, visual and multilingual, comparing the effects of processes in multiple languages. The activity, open to teams from all over the world, seeks to encourage methods that allow access to large amounts of images and data, used in scenarios such as information retrieval, machine learning, deep learning, data mining and image and video processing, all this through challenges that involve factors such as multimodality and interactive searches. For Denis Parra, an academic at the DCC UC and part of the Millennium Institute Fundamento de los Datos (IMFD), this result is of particular merit as it is his first activity as a team:

"We put together the team in January of this year, with almost pure undergraduate students and only two who have just started the master's degree, all with very little experience in the subject, but with two" trainers "in Pablo Pino (Masters degree in DCC UC) and Pablo Messina (Ph.D. DCC UC). Several left with homework for their vacations and continued studying the topics and practicing the use of Pytorch and other deep learning tools during the holidays. Since March they have had a heavy course load in the semester And yet we get together in March and April on Fridays, sharing data analysis, discussing strategies, debugging code, collaborating, etc. It is a very young team, we will probably participate again next year and we will surely do better. "Said Parra." I emphasize that it was key to have a server with 3 GPUs that I bought with the Fondecyt that I won in 2019. "he added.

The  IALab / HAIvis UC was composed by:

  • Académicos:  Hans Lobel y Denis Parra
  • Senior Coaches:Pablo Pino (Masters) y Pablo Messina (PhD)
  • Visual Question Answering category: Ricardo Schilling (undergraduate)
  • Image Captioning - Caption detection category:Vicente Castro (undergraduate)
  • Image Captioning - Concept detection category: Greg Schuit y Vicente Castro (undergraduate)
  • Tuberculosis CT analysis category: Ria Deane, Daniel Florea, José Miguel Quintana(undergraduate)
All Posts
×

Almost done…

We just sent you an email. Please click the link in the email to confirm your subscription!

OK