Time to team-up deadline

0 / 0

participants / team


Topic provider

The National Museum of Taiwan Literature (NMTL) is the only museum of literature at the national level in Taiwan. The operational guidelines of the NMTL are to completely collect works on the history of Taiwan literature and make full use of them. The NMTL has more than 100,000 important rare books, writers’ manuscripts, and letters and articles on the history of Taiwan literature. The NMTL has the heavy responsibility of collecting, researching, publishing, displaying and promoting Taiwan literature. Along with the spaces of Library and Literary Wonderland to provide diversified services.
The Executive Yuan passed the third reading about the National Museum of Taiwan Literature Organization Act in 2021, designed to raise the administrative level of the museums from fourth-level to third-level agencies, which would benefit the NMTL’s work in the preservation, research and promotion of Taiwanese literature. In the future, the NMTL will strengthen the cultivation of preservation professionals and the capabilities of literary collection, and also to expand the literature-related space and link literary families in Taiwan.
National Museum of Taiwan Literature:
NMTL's Archive Select:
Digital Archive of Collection:
Virtual Museum of Taiwan Literature:


The National Museum of Taiwan Literature (NMTL) is representative of museums housing collections of Taiwanese literature in terms of both quality and quantity. NMTL collects representative, important works and references in the history of Taiwanese literature. Examining and keeping a record of said collections on a regular basis enables NMTL staff to effectively keep track of their current status. NMTL has engaged in the task of examining and keeping a record of its collections for more than a decade and accumulated fruitful results. However, it is very time-consuming to examine the massive number of collections of over 110,000 works one by one in a manual manner. Consequently, this topic focuses on developing AI-based technologies for deterioration detection of collections . In the future, a database of deteriorated literary collections will be created. This database will not only help NMTL staff improve their efficiency in examining NMTL’s collections, but it will also support comparisons of the collections before and after exhibitions, while saving on manpower and time required for traditional photographing and examination techniques.

Those engaging in the fields of data analysis and science are invited to attend this event to provide training-related AI models, so as to help detect positions and types of deterioration of literary collections via the provided image data. It is hoped that this event can benefit NMTL in examining its collections.


First place: USD$2,000 (tax inclusive)

Second place: USD$1,000 (tax inclusive)

Third place: USD$500 (tax inclusive)


Activity time

The time of this event is Indian Standard time (UTC+5.5). The agenda is as follows:

Time (YYYY/MM/DD)Events
2021/12/06Registration starts
2021/12/17Available for download
2022/01/03Online submission available
2022/03/18Online submission deadline
2022/03/19Private Leaderboard announcement
2022/03/25Reports submission deadline
2022/03/31Prize Winners announcement


Evaluation Criteria

Mean Average Precision (mAP) [1] will be adopted for this topic’s data evaluation metrics; the Intersection over Union (IoU) [2] threshold is 0.5. When the IoU of predicted and labelled bounding boxes containing objects is greater than 0.5, it means True Positive (TP). Conversely, when the IoU is less than 0.5, it means False Positive (FP). The measured values will be used to calculate precision. The AP score of each object will be evaluated via the system. The average of the AP scores of the five types of objects that have been deteriorated will be calculated to obtain mAP evaluation values. Participants will be ranked in accordance with said values. COCO API will be adopted for the calculation of mAP evaluation values.

[1] Average Precision (AP):

[2] intersection over union (IoU):



  • The evaluation will be based on the final uploaded result. If multiple participants got the same evaluation scores, the time of uploading would determine the ranking.
  • The maximum number of uploading times is 5 times per day.
  • Teaming up with others is allowed. One person can only have one account. Repeat participation is prohibited. Violators will be disqualified from participation upon verification.
  • Participating teams cannot partake in the same competition using multiple accounts. Violators will be disqualified from participation. If a member of a participating team uploads a file using his/her personal account, the team will be deemed to have participated in the competition using multiple accounts.
  • Each person can only register for a team. He/she cannot change to another team once he/she joins the team. However, up to five people are allowed to join the team.
  • When using external data sets, participants should avoid using future data as the basis of the prediction results, and must state source of data sets in the forum for references.
  • All the data, techniques and source codes are original work of participants or are used by permission complying with laws and regulations. If any third-party claims their intellectual property rights or other rights and interests are being violated, the participant will need to handle the disputes personally. If any participants violate intellectual property rights, they will be disqualified, and shall bear legal responsibilities.
  • All the achievements and their IPRs (intellectual property rights) belong to the participants, and the Copyright License Agreements, patent applications, technology transfer and equity distributions of them, should be in accordance with the relevant Laws and Regulations.
  • There should be no answer discussions between different accounts, or it will be considered as cheating.
  • If there is cheating or fraud during the activity, the participant that cheats will be disqualified from the activity and the vacancy would be filled up by other participants in the ranking order.
  • After uploading, the answers of test data would be divided into two parts to calculate the score,
    • a. Before the deadline of the activity:
      The system refers to partial answers of test data to examine and calculate the score. The result will be posted on Public Leaderboard. This data accounts for 60% of the entire test data.
    • b. After the deadline of the activity: The system refers to the remaining test data (100%) to examine and calculate the score. The result will be posted on Private Leaderboard as reference for the final score and ranking.
  • Should there be disputes, the organizers reserve the right to the final decision.