ImageCLEF

23/02/2024

& https://lnkd.in/d5Qk4wWD -- generating real-world medical images, i.e., examining the hypothesis that GANs are generating medical images that contain certain "fingerprints" of the real images used for generative network training identify training data “fingerprints”, detecting generative models' “fingerprints” -- if the hypothesis is correct, artificial biomedical images may be subject to the same sharing and usage limitations as real sensitive medical data. 06-05-2024 are to be presented at Conference and Labs of the Evaluation Forum, 9-12 September 2024, Grenoble, France AI4Media Universitatea POLITEHNICA din București HES-SO Valais Wallis.

23/02/2024

& https://lnkd.in/dwtAw9QG -- Europeana.eu: with over 53M records, the single search bar that served as the main access point was identified as a bottleneck by many users. Now users can explore over 60 curated digital exhibitions, countless galleries and blog posts. While there is a system in place to individual items given a query item, the recommendations for editorials are done now only manually. given a list of items/editorials, use AI techniques to provide a list of recommended items/editorials. 6-5-2024 are to be presented at Conference and Labs of the Evaluation Forum, 9-12 September 2024, Grenoble, France AI4Media Universitatea POLITEHNICA din București .

05/02/2024

ImageCLEFmedical Caption in its 8th edition is open for registration and the training/validation dataset release is soon to come! Have a look at the attached Call for Participation and visit the website (https://www.imageclef.org/2024/medical/caption) for more information!

Interpreting and summarizing the insights gained from medical images such as radiology output is a time-consuming task that involves highly trained experts and often represents a bottleneck in clinical diagnosis pipelines.

15/01/2024

lab https://www.imageclef.org/2024 -- it evaluates the technologies for , , and of data, with its main objective in providing to large collections of multimodal data for multiple usage scenarios and domains.

The 2024 tasks are: with image captioning, image synthesis and optimal prompt generation, multimodal and generative telemedicine, medical GANs; heritage recommendation; ToPicto pictogram translation of text/speech; and image retrieval for arguments (joint task).

Results are to be presented at the annual 2024 Conference and Labs of the Evaluation Forum, in Grenoble, France, 9-12 September 2024.

https://ai4media-bench.aimultimedialab.ro/competi.../public/ via the AI4Media benchmarking platform (based on Codalab).

Universitatea POLITEHNICA din București HES-SO

09/03/2023

1st Cultural Heritage Content-based Task https://www.imageclef.org/2023/recommending:

Media archives have not only exponentially increased in size, but now hold contents in various modalities (e.g., video, image, text). Even when structured metadata is available it is still difficult to discover the contents and allow users to navigate multiperspectivity. Content-based recommendation systems can help but there is limited understanding how well these perform and how relevant they are for the end-users. Moreover, the system used so far have not addressed the new user requirements of more transparency and explainability of the algorithms used.

The task requires participants to devise recommendation methods and systems, apply them in the supplied data set gathered from Europeana and provide a series of recommendations in two scenarios: (i) given a list of items provide a list of recommended items; (ii) given an editorial (Europeana blog or gallery) provide a list of recommended editorials.

submission: May 10, 2023
Working notes submission: June 5, 2023
CLEF 2023 conference: September 18-21, Thessaloniki, Greece

Universitatea POLITEHNICA din București AIMultimediaLab AI4Media ImageCLEF Europeana.eu

09/03/2023

1st Cultural Heritage Content-based Task https://www.imageclef.org/2023/recommending:

Media archives have not only exponentially increased in size, but now hold contents in various modalities (e.g., video, image, text). Even when structured metadata is available it is still difficult to discover the contents and allow users to navigate multiperspectivity. Content-based recommendation systems can help but there is limited understanding how well these perform and how relevant they are for the end-users. Moreover, the system used so far have not addressed the new user requirements of more transparency and explainability of the algorithms used.

The task requires participants to devise recommendation methods and systems, apply them in the supplied data set gathered from Europeana and provide a series of recommendations in two scenarios: (i) given a list of items provide a list of recommended items; (ii) given an editorial (Europeana blog or gallery) provide a list of recommended editorials.

submission: May 10, 2023
Working notes submission: June 5, 2023
CLEF 2023 conference: September 18-21, Thessaloniki, Greece

Universitatea POLITEHNICA din București AIMultimediaLab AI4Media ImageCLEF Europeana.eu

15/01/2023

Multimedia Retrieval in CLEF https://lnkd.in/dWPcdbzb:
(new) medical dialogue topic classification and summarization https://lnkd.in/dXXKFmY8
(new) visual question/question location answering, and generation in colonoscopy images https://lnkd.in/ddiTvtfj
(new) traceability of training data in synthetic medical image generation with GANs https://lnkd.in/dgeKRtBW
(7th edition) concept detection and caption prediction https://lnkd.in/deVKs4kP
(new) meaningful and diverse recommendations of articles and editorials from Europeana data https://lnkd.in/dMve8N6V
(3rd edition) automatic classification of photographic social media user profiles in unintended scenarios https://lnkd.in/diyS6aUG
(2nd edition) late fusion mechanisms and ensembling in interestingness prediction, search results diversification, and image captioning https://lnkd.in/d2PVWZri
is due May 10, 2023 --- AI4Media Universitatea POLITEHNICA din București AIMultimediaLab HES-SO.

08/03/2022

special session on "Learning from scarce data challenges in the media domain" https://cbmi2022.org/call-for-special-session-papers/ -from-scarce-data @ 19th International Conference on Content-based Multimedia Indexing (CBMI 2022) AI4Media. Papers are due April 10, 2022.

Deep learning-based algorithms for multimedia content analysis need a large amount of annotated data for effective training, e.g., for image classification on the ImageNet dataset, each class comprises several thousand annotated samples. Having a dataset of insufficient size for training usually leads to a model which is prone to overfitting and performs poorly in practice. But in many real-world applications in multimedia content analysis, it is not possible or not viable to gather and annotate such a large training data. This may be due to the prohibitive cost of human annotation, ownership/copyright issues of the data, or simply not having enough media content of a certain kind available. To address this issue, a lot of research has been performed in recent years on learning from scarce data/learning from limited data. There are a variety of ways to work around the problem of data scarcity like using transfer learning, domain transfer or few-shot learning.

10/02/2022

Machine Learning System Fusion Challenge https://www.imageclef.org/2022/fusion:

While deep neural networks have proven their predictive power in many tasks, there are still many areas where a single deep learning network is not enough for attaining high precision. Late fusion (ensembling, decision-level fusion) represents one of the approaches that researchers employ to increase the performance of single-system approaches.

The participants will receive a data set of real machine learning systems and are expected to develop fusion mechanisms that would allow to combine them into a super-system yielding superior performance compared to the highest performing individual system. Two real-world use-cases are provided: (i) a task via systems for predicting visual interestingness, and (ii) a task via systems for image search results diversification.

https://www.imageclef.org/2022
Run submission: May 6, 2022
Working notes submission: May 27, 2022
CLEF 2022 conference: September 5-8, Bologna, Italy

AI4Media Universitatea POLITEHNICA din București ImageCLEF

29/01/2022

Unveiling Real-Life Effects of Online Photo Sharing https://lnkd.in/dJh-VTJP :

Images constitute a large part of the content shared on social networks. Their disclosure is often related to a particular context and users are often unaware of the fact that, depending on their privacy status, images can be accessible to third parties and be used for purposes which were initially unforeseen. For instance, it is common practice for employers to search information about their future employees online.

The objective of the task is to automatically score user photographic profiles via AI in a series of situations with strong impact on her/his life, e.g., search for a bank loan, an accommodation, a job as waitress/waiter, a job in IT. This is the second edition of the task. To train the algorithms, a data set of 1,000 user profiles with 100 photos per profile was created and annotated with an appeal score for a series of real-life situations via crowdsourcing.

https://lnkd.in/dc9WtyUa
Run submission: May 6, 2022
Working notes submission: May 27, 2022
CLEF 2022 conference: September 5-8, Bologna, Italy

CEA LIST AI4Media Universitatea POLITEHNICA din București ImageCLEF

18/01/2022

@ https://www.imageclef.org/2022 : >> ImageCLEFcoral (4th edition) coral reef and , and image pixel-wise https://www.imageclef.org/2022/coral; >> ImageCLEFmedical (4th edition) automatic with , and analysis of https://www.imageclef.org/2022/medical; >> ImageCLEFaware (2nd edition) development of algorithms which raise the users’ about real-life impact of https://www.imageclef.org/2022/aware; >> ImageCLEFfusion (new) develop mechanisms and generate predictions with significantly higher performance than the individual systems https://www.imageclef.org/2022/fusion.
***Run submission is due May 6, 2022*** ImageCLEF AI4Media Universitatea POLITEHNICA din București HES-SO La Rochelle Université.

17/01/2022

call-for-papers ACM on against @ ACM ICMR 2022, Newark, NJ, USA, June 27-30, 2022 https://mad2022.aimultimedialab.ro/ --- Disinformation campaigns are increasingly powered by advanced AI techniques and a lot of effort was put into the detection of fake content. While important, this is only a piece of the puzzle if one wants to address the phenomenon in a comprehensive manner. Whether a piece of information is considered fake or true often depends on the temporal and cultural contexts in which it is interpreted. This is for instance the case for scientific knowledge, which evolves at a fast pace, and whose usage in mainstream content should be updated accordingly --- our workshop to discuss multimedia AI methods to fight the ever growing online disinformation campaigns --- is due: February 17, 2022 Universitatea POLITEHNICA din București Εθνικό Κέντρο Έρευνας και Τεχνολογικής Ανάπτυξης CEA LIST Fraunhofer IDMT AI4Media ACM - Association for Computing Machinery.