ICME 2026 APS Challenge
Challenge Description
Understanding ancient diets is essential for reconstructing subsistence strategies and human–environment interactions. Plant seeds are key archaeobotanical evidence, yet their identification still relies on time-consuming expert analysis. Existing datasets mainly contain well-preserved samples and fail to reflect real archaeological conditions.
To address this gap, the APS Challenge extends the original APS dataset by incorporating severely damaged and carbonized seed images. Participants are invited to develop robust fine-grained classification models that handle high inter-class similarity and large intra-class variation under realistic conditions.
📅 Important Dates (All deadlines are at 23:59, UTC)
- Registration Open & Training/Validation Data Release 2026-03-01
- Testing Data Release 2026-04-01
- Challenge Result Submission Deadline 2026-04-10
- Weight File & Code Submission Deadline 2026-04-15
- Final Decisions Announcement 2026-04-20
- Challenge Technical Report Submission Deadline 2026-04-25
Dataset
The APS Challenge dataset extends the original APS (Ancient Plant Seeds) dataset with severely damaged and carbonized seed images, providing a more realistic representation of archaeological conditions. The dataset includes various seed categories with fine-grained annotations to facilitate the development of robust classification models.
Figure 1: Comparison of the size and damage of ancient plant seeds and the condition of modern seeds. We divide the seeds into four sizes, where each has three columns of seed images. The first column shows the true seed size under 1.6× magnification, the second column shows seeds with pronounced differences caused by uncontrollable factors, and the third column shows the corresponding modern seeds.
📁 Dataset Structure
Train
Validation
Test
Rules
📝 Submission Guidelines
Participants are required to submit prediction results through the official challenge website (such as Codabench, link to be published later).
Submissions should be provided as a result.txt file, where each line follows the format:
image name + predicted label.
Challenge Phases
Phase I: Participants submit prediction results on the validation set to obtain evaluation metrics and A-list real-time rankings.
Phase II: Participants submit prediction results on the test set to receive evaluation metrics and B-list real-time rankings.
Each team is limited to a maximum of five submissions per day.
📊 Evaluation Criteria
As this is a multi-class classification task, performance will be evaluated using Accuracy, Precision, Recall, and F1-score. Accuracy and F1-score will serve as the primary ranking metric.
Evaluation Metrics
Where: TP = True Positive, TN = True Negative, FP = False Positive, FN = False Negative
Leaderboard System
- A-rank (Initial Leaderboard): Based on results on the validation set
- B-rank (Final Leaderboard): Based on evaluation on the test set
The B-rank leaderboard will be used as the final ranking of the challenge.
ℹ️ Notes
- The participates can form their own teams from different organizations and the number of participants is not limited. But one person can only join one team.
- The participates are NOT allowed to use external data for either training or validation.
- Participants were not allowed to use additional information manually labeled on the training or validation datasets to identify the target label of the challenge.
- The provided dataset can only be used for academic purposes. By using this dataset and related software, you agree to cite our dataset and baseline paper.
- The top three participants will receive a certificate based on the ranking of Testing Data as the final list.
- Contestants were allowed to use methods other than the baseline model.
Baseline Method (APSNet)
We provide a baseline model, APSNet, while allowing participants to adopt any methods beyond the baseline. Our previous work, which has been publicly released on arXiv, is adopted as the baseline method for this challenge and also serves as the source of the APS dataset. The proposed APSNet demonstrates competitive performance on fine-grained ancient plant seed classification. The source code of the baseline method is publicly available at: https://github.com (Come Soon). Participants are encouraged to cite our baseline paper if it is relevant to their research.
⭐ BibTeX
Come SoonOrganizers
🚩 Runmin Cong
Professor, Shandong University
Director of MVP Group
🚩 Can Wang
Professor, Shandong University
Member of Archaeological Society of China
👥 Wei Zhang
Professor, Shandong University
Vice Dean of School of Control Science and Engineering
👥 Fen Wang
Professor, Shandong University
Dean of School of Archaeology
👥 Sam Kwong
Professor and Vice President, Lingnan University
🚀 Join the Challenge!
Advance the state-of-the-art in archaeological artifact identification and contribute to the preservation of human heritage.
📧 Contact Us
For questions about the challenge, please contact the organizers at: apschallenge@163.com
