Wing / boxing-moves Public

Target: Arduino Nano 33 BLE Sense (Cortex-M4F 64MHz) Clone this project

Welcome to Edge Impulse, the largest community of edge AI developers!
This is a public Edge Impulse project, use the navigation bar to see all data and models in this project; or clone to retrain or deploy to any edge device.

boxing-moves

Accelerometer

About this project

This project will collect boxing moves (jab, cross, hook and uppercut) and subsequently train a model for motion recognition.

Data (150 samples)

View all

KHook.2tubtdk2

ID: 84895905

Label: Hook

Sample length: 10s

RJab.2u1j6jn1

ID: 84895838

Label: Jab

Sample length: 10s

KUppercut.2tubr8s7

ID: 84895904

Label: Uppercut

Sample length: 10s

MHook.2u1mq9ku

ID: 84895821

Label: Hook

Sample length: 10s

LeftUppercut.2tv0811n

ID: 84895857

Label: Uppercut

Sample length: 10s

JumpCross.2tulvf74

ID: 84895879

Label: Cross

Sample length: 10s

Uppercut.2tuk2aro

ID: 84895885

Label: Uppercut

Sample length: 10s

MMAJab.2tvgdnql

ID: 84895847

Label: Jab

Sample length: 10s

Model accuracy (Unoptimized (float32))

Validation set

88.5%

Test set

87.9%

On-device performance (Arduino Nano 33 BLE Sense (Cortex-M4F 64MHz), Quantized (int8))

Latency

12 ms.

Peak RAM usage

5.1K

Flash usage

19.9K

Run this model

On any device

Clone project

Dataset summary

Data collected

25m 0s

Sensors

accX, accY, accZ @ 100Hz

Labels

Cross, Hook, Jab, Ready, Uppercut

Project info

Project ID	86666
Project version	4
License	No license attached
No. of views	14,946
No. of clones	42