Biomedisa's deep learning capabilities enable automated segmentation, facilitating the comparative analysis of distinct objects such as insect brains. Additionally, Biomedisa's functionality extends to segmenting repetitive structures within large volumes, spanning from cells to rock particles to roots in soil, all without the need for image resizing.
In this talk, I demonstrate Biomedisa's patch-based deep learning approach to segment structures within volumes of up to 2,500 x 2,500 x 10,000 voxels. Biomedisa successfully identifies plant roots in fertilized soil and separates thousands of particles from a crushed rock scan by predicting particle boundaries and separating connected regions.
Furthermore, I introduce the software HEDI, a biomechanical approach to incisional hernia repair, and show how Biomedisa’s results and visualizations can assist surgeons in pre-operative planning, with promising initial clinical results.
Lastly, I will present the ANTSCAN project, a pilot initiative focusing on the digitization of insect morphological diversity, aimed at capturing phenotypes across the ant tree of life. Biomedisa's online platform supports this initiative by providing access to tomographic data, comprehensive metadata, and 3D morphological insights for over 2208 scanned ants from 850 species, enabling the global scientific community to collaboratively analyze and contribute to this extensive database.
Room:
Physics Auditorium