Hi tls,
Back on the farm, we used to count the legs and divide by four.
This is a proof-of-concept arrangement, so not optimized for commercial exploitation. It is at this stage an algorithm running in Raspberry Pi.
The heat sink indicates the processor is generating a bit of heat. I'm not sure if they could dispense with the heat sink if they had Akida, but they could use a smaller one, which would lighten the load for the drone, and, as we know, Akida would extend the battery charge.
Apparently Edge impulse does a lot of this proof-of-concept modelling:
https://www.edgeimpulse.com/blog/getting-more-cycles-per-second-with-fomo
Note that this system does not appear to have one-shot learning, and needs a significant training database.
Basically they are using Edge Impulse's FOMO (Faster Objects, More Objects) algorithm:
https://www.edgeimpulse.com/blog/announcing-fomo-faster-objects-more-objects
It appears that there are synergies in using FOMO and Akida.
I hadn't previously looked into EI's FOMO, but the above link provides an excellent introduction, and the synergies become apparent.
FOMO uses a form of CNN and has significant limitations.
FOMO uses a form of CNN and has significant limitations.
Maybe EI will win a prize for FOMO at the Vision trade fair in Stuttgart.
Last time Prophesee won (
https://www.imveurope.com/analysis-opinion/neuromorphic-pioneer-recognised-vision-stuttgart).
"Trend topics and innovations in focus
The vision presents the lecture program
11. September 2022
3 minutes reading time
At the Industrial Vision Days, visitors can expect visionaries from the image processing industry
Photo: Landesmesse Stuttgart
The image processing industry is developing dynamically and is becoming increasingly important. Innovative products are constantly conquering new markets and offering solutions for different industries.
Trending topics and innovations in the image processing industry will be presented and discussed at the Vision from October 4th to 6th, 2022 in Stuttgart. Special highlights in this context are the Industrial Vision Days and the award ceremony.
The Industrial Vision Days are organized by VDMA Machine Vision together with Messe Stuttgart. The focus of the forum:
Visionaries from the image processing industry.
Mark Williamson, Chairman of the VDMA Machine Vision department: "The Industrial Vision Days will again offer many exciting lectures from all areas of image processing, be it camera technology, robot vision and 3D vision, software and AI, optics and lighting, hyperspectral image processing, image processing standards and new applications. The approximately 50 company presentations, 16 start-up pitches and the presentations of the five finalists of the Vision Award show how innovative the image processing industry is. A look at the Industrial Vision Days program makes it clear that deep learning is still the hot topic. Many of the start-up pitches promise an innovative leap through deep learning. I'm excited and will be happy to be convinced as a jury member in the three start-up pitch sessions."
The forum is an important stage for the participating companies. They have the opportunity to present groundbreaking innovations in front of a large specialist audience.
Martin Grzymek, Director Sales Europe at Teledyne GmbH - Dalsa Division, underlines the high relevance of the Industrial Vision Days: "The Industrial Vision Days are an indispensable platform that enables Teledyne to establish personal contacts with users. In short, a very welcome opportunity for us to present the latest developments and learn from users how their demanding applications are driving innovation in machine vision.”
The lecture program of the Industrial Vision Days is already complete
available online.
An important part of the program is the Vision Award, which will again be presented on the forum stage. For the 25th time, the best innovations in the field of image processing will be chosen this year. “We received a total of
61 submissions. This corresponds to an
increase of almost 40 percent compared to the previous year. That's an impressive number," says Florian Niethammer, Head of Trade Fairs and Events at Messe Stuttgart.
The
award jury drew up a shortlist of the most interesting developments from all submissions. “The trend topics of image processing were found among the submissions for the 2022 award: embedded vision,
artificial intelligence, components, vision toolboxes as well as algorithmic and application-specific work. The high quality of the submissions this year deserves special mention. As a result, the five companies Brighter AI Technologies GmbH,
Edge Impulse, Kitov.ai, Saccade Vision Ltd. and SWIR Vision Systems Inc. will present their innovations live at the Vision,” summarizes Christian Ripperda, Managing Director of Interroll Innovation GmbH and a member of the jury for the award."
https://epp.industrie.de/events-termine/die-vision-praesentiert-das-vortragsprogramm/
"By Mihajlo Raljic, Edge Impulse
Description of the innovation:
A new machine learning technique developed by researchers at Edge Impulse, a platform for creating ML models for the edge, makes it possible to run real-time object detection on devices with very small computation and memory capacity. Called Faster Objects, More Objects (FOMO), the new deep learning architecture can unlock new computer vision applications.
Most object-detection deep learning models have memory and computation requirements that are beyond the capacity of small processors. FOMO, on the other hand, only requires several hundred kilobytes of memory, which makes it a great technique for TinyML, a subfield of machine learning focused on running ML models on microcontrollers and other memory-constrained devices that have limited or no internet connectivity. TinyML has made great progress in image classification, where the machine learning model must only predict the presence of a certain type of object in an image. On the other hand, object detection requires the model to identify more than object as well as the bounding box of each instance. Image classification is very useful for many applications. For example, a security camera can use TinyML image classification to determine whether there's a person in the frame or not. However, much more can be done.
Technical details and advantages of the innovation:
...
...
...
When will the innovation be available?
already available"
https://www.messe-stuttgart.de/visi...ortlist-der-besten-einreichungen#edge-impulse