Benchmarks and Competition
A - Tile:
The competition will be held as part of the International Workshop on Photogrammetric and computer vision techniques for video surveillance, biometrics, and biomedicine - PSBB23 in Moscow. The goal is to challenge the morphology of teeth.
We will provide 224 3D models of teeth, representing paleo-anthropological findings. The type is classified by its morphology. The description of the classification scheme will be provided applicants (Figure 1).
Figure 1. Upper. Description of tooth type classification. Below. Example of the dataset.
The tasks of the contest are (1) automatic tooth type classification; (2) automatic morphological analysis, and textual annotation. For this point, the rules will be provided in a few weeks.
Through this competition, we want the participating research teams to find the best solution. Therefore, the prize will be awarded to the best numerical solution at the end of the PSBB23.
B - Tile:
The test is from the Viedma glacier, Southern Patagonia Icefield, Argentina (49° 31′ S, 72° 59′ W) located at Los Glaciares National Park. This site is an important calving glacier in the region covering an area of 900 km2. The behavior of the glacier at the terminus is of particular interest because it sheds a large amount of ice floes by losing ice mass. To support the field image acquisition of the time-lapse imagery, an integrated data acquisition system was built based on a NIKON D3 camera (C2) calibrated by Rollei Metric pixel size: 8.5 μm, objective focal length: 35 mm, and FOV: 62°. The system was powered by one 12V/7Ah lead acid battery, charged by two 38W solar panels. The camera with the supporting electronic system was protected by a waterproof box with a visor to reduce reflections. The image acquisition system was installed on a rigid metal structure, fixed to outcrops of the south margin of the Viedma Glacier. We will provide 800 images approximately (multitemporal raw data) for the front of Viedma Glacier (see Figure 2). A description of the dataset will be provided to applicants.
Figure 2. Example of the pictures of the dataset.
The tasks of the contest are (1) multi-temporal scaling model; (2) multitemporal model of the front geometry, and errors impact. The time the teams will have to process the data will be 200 days. For the application of the working teams, they should submit a mini proposal as a workflow with the proposed methodology to be used. Once this stage is completed, the dataset will be released to the applicant groups. We highlight that the dataset is the property of the Andean Geomatics Lab (IANIGLA-CONICET-ARGENTINA).
Through this competition, we want the participating research teams to find the best methodological and numerical solution. Therefore, the prize will be awarded to the best methodological solution at the website of the ISPRS.