We are glad to announce the 1st edition of the International Competition on Few-Shot and Many-Shot Layout Segmentation of Ancient Manuscripts, in conjuction with the 18th International Conference on Document Analysis and Recognition ICDAR 2024.

Layout analysis is a critical aspect of Document Image Analysis, particularly when it comes to ancient manuscripts. It serves as a foundational step in streamlining subsequent tasks such as optical character recognition and automated transcription. However, one key challenge in this context is represented by the lack of available ground truths as they are extremely time-consuming to produce. Nevertheless, numerous approaches addressing this challenge heavily lean towards a fully supervised learning paradigm, which is difficult to use in real-world scenarios. For this reason, with this competition, we propose the challenge of addressing this task with a few-shot learning approach, involving the use of only three images for training. The competition dataset comprises four distinct ancient manuscripts, presenting heterogeneous layout structures, levels of degradation, and languages used. This diversity adds intrigue and complexity to the challenge. In addition, we have also created the opportunity to participate in the competition with the traditional many-shot learning approach, involving the use of multiple images for training.

Tasks

The Tasks we propose for this competition are two:

• Task 1: Few-Shot Layout Segmentation. This Track, in our opinion, is the most ambitious and interesting. It involves creating an effective document layout segmentation system using only three images for each manuscript. Ten additional images, with the corresponding ground truth, are provided for validation only. It is strictly forbidden to use these additional images as a training set!
• Task 2: Many-Shot Layout Segmentation. Conversely, this task is more conventional. Participants are required to create a layout segmentation system using 20 images per manuscript along with their corresponding ground truth, divided into training and validation sets.

Important Dates

• January 10, 2024: Open registration to competition
• January 15, 2024: Beginning of the competition, Track 1
• March 3, 2024: Deadline of Track 1 (short report + code)
• March 11, 2024: Deadline of Track 1 (short report + code)
• March 4, 2024: Beginning of the competition, Track 2
• March 12, 2024: Beginning of the competition, Track 2
• March 31, 2024: Deadline of Track 2 (short report + code)
• April 8, 2024: Deadline of Track 2 (short report + code)
• April 15, 2024: Deadline of Track 2 (short report + code)
• May 31, 2024: Competition reports submission deadline
• Announcement of the winners

Prizes and Awards

For each Track, we plan to nominate a winner based on the evaluation process of the proposed system. The winners of the two tracks will receive a cash prize of 300 EUR each sponsored by CVPL – Italian Association for Computer Vision, Pattern Recognition and Machine Learning, IAPR Italian chapter. The prize for the winners will be awarded only if the competition report will be accepted for publication on ICDAR proceedings (call for competition). In particular, the competition could be rejected if the number of participants is not enough to draw meaningful conclusions.

News :

• April 8, 2024, Deadline extension for submission Track 2 to April 15th
• March 4, 2024, Deadline extension for submission Track 1 to March 11th
• March 4, 2024, update evaluation code
• March 1, 2024, added new delivery instructions in the Data & Tools section

Winners

Track 1: Few-Shot Layout Segmentation

Position	Team	Average Iou
1	CV-Group, Computer Vision Group of Pattern Recognition Lab of Friedrich- Alexander-Universitat Erlangen-Nurnberg	0.784
2	VAI-OCR, Artificial Intelligence department of Viettel AI from Vietnam	0.700
3	CNKI, China National Knowledge Infrastructure, Large Model and Future Technology R&D Department	0.659
4	L3i++, L3i laboratory at University of La Rochelle, France	0.611
5	PRAIG-UA, Pattern Recognition and Artificial Intelligence Group of the University of Alicante, Spain	0.465

Task 2: Many-Shot Layout Segmentation

Position	Team	Average Iou
1	CV-Group, Computer Vision Group of Pattern Recognition Lab of Friedrich- Alexander-Universitat Erlangen-Nurnberg	0.834
2	CNKI, China National Knowledge Infrastructure, Large Model and Future Technology R&D Department	0.778
3	VAI-OCR, Artificial Intelligence department of Viettel AI from Vietnam	0.707
4	PRAIG-UA, Pattern Recognition and Artificial Intelligence Group of the University of Alicante, Spain	0.475