TinyML on Arduino Nano 33 BLE Sense - Fruit Identification

RE: TinyML on Arduino Nano 33 BLE Sense - Fruit Identification

neilk — Wed, 20 Jan 2021 11:34:57 GMT

Jon, I'm also late [View:/resized-image/__size/16x16/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-4e366a08-37d9-4caa-a73c-0347705d1903/contentimage_5F00_734.png:16:16]

Very impressive results.

Neil

RE: TinyML on Arduino Nano 33 BLE Sense - Fruit Identification

hugohu — Tue, 19 Jan 2021 23:37:06 GMT

Nice blog. I look forward to the day I get a 33 Sense and follow this. Thanks for writing [View:/resized-image/__size/16x16/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-4e366a08-37d9-4caa-a73c-0347705d1903/contentimage_5F00_2516.png:16:16]

RE: TinyML on Arduino Nano 33 BLE Sense - Fruit Identification

14rhb — Tue, 19 Jan 2021 20:48:59 GMT

A bit late to the party, but now with an Arduino 33 Sense in my hand, I search the internet and find this great article by you [mention:d532f5dec7024823bdf402d3f01f1359:e9ed411860ed4f2ba0265705b8793d05] - so thank you for all this detail to get me started [View:/resized-image/__size/16x16/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-4e366a08-37d9-4caa-a73c-0347705d1903/contentimage_5F00_1.png:16:16]

RE: TinyML on Arduino Nano 33 BLE Sense - Fruit Identification

hyemi — Wed, 01 Jul 2020 07:36:06 GMT

Thank for very useful posting.

Why tensor_arena size is 8KB which is much more than model size(1768 byte)?

Thank you in advance.

RE: TinyML on Arduino Nano 33 BLE Sense - Fruit Identification

ankur608 — Wed, 08 Apr 2020 12:53:32 GMT

Great article. It was fun to see the variation(a bell pepper) in this case.

RE: TinyML on Arduino Nano 33 BLE Sense - Fruit Identification

dubbie — Wed, 08 Apr 2020 11:34:40 GMT

Jon,

Another very interesting example of machine learning. I liked the idea of getting emoji's against each category although I'm not sure how easy it would be to extend to other topics. Maybe you could try recognising sweets from a tin of Quality Street, or similar.

It would be relatively easy to connect a simple sensor that provided some 2D data, such as a line of 8 optical detector connected to 8 inputs on the Nano Sense. Then you could recognise shapes as well as colours.

My request for your next experiment is to use BLE as I will shortly be planning (hoping?) to use BLE to communication with my current Nano project. I used to know a bit about Bluetooth, but this BLE seems to be completely different so I have a big learning curve to follow. I just want a Bluetooth serial communication link.

Dubbie