Real-time alerts from AI-enabled camera traps using the Iridium satellite network: a case-study in Gabon, Central Africa
10 November, 2021Read Paper
1 week from concept to working prototype
See more than human auditors are capable of
Automate tasks that once took many hours
25 cm maximum satellite image resolution
Mbaza is used in the national parks of Gabon to help protect and monitor their pristine ecosystems. It is powered by image classification models, allowing the ecologists and park rangers to tap into insights from the data collected by camera traps. Using their existing equipment, they can process the data within a day. Without the tool, this typically took a month. Read more
In this research and development project, we partnered with a renowned motion capture studio and experts in human motion analysis. We developed algorithms generating human motion while controlling the emotional flavor of it. We achieved results reaching beyond the state of the art, in both speed and accuracy. Paper in preparation.
The vitality of plankton is a crucial window into the health of the Arctic ocean. We have developed methods of assessing the nutritive value of plankton using Image Segmentation, which opens the possibility of monitoring the Arctic oceans’ ecosystem on a large scale. This is particularly important in the era of accelerated shifts caused by the climate crisis. Read more
A Machine Learning engine of a platform allowing swift design of 3D characters based on a 2D vision by an artist. The process of creating 3D characters following the creators' ideas is shortened from weeks to minutes. We combined various computer vision techniques and worked closely with graphic designers, 3D sculptors, and animators. Read more
With smart cameras, it becomes feasible to monitor remote ecosystems in real-time. This makes it possible to, e.g., alert farmers in Central Africa that Elephants are about to intrude in their fields before it happens. We have prepared Computer Vision models which were utilized in the smart cameras tested in Central Africa. Read more
I can attest to exceptionally professional and effective workflow and great project management at Appsilon.
[They're able to] understand project needs and to develop innovative solutions that we had not previously envisioned.
We’re impressed that they are results-oriented.
Their genuine curiosity and impressive ability to adapt to our needs were outstanding!
We were impressed with the professionalism of the team and their project management style.
Track, reproduce, and share your ML models with DVC for data versioning
Actions taken and lessons learned at a 24-hour hackathon dealing with fly-tipping problems.