Deep Learning Inference Demo: CIFAR-10 Image Classification on PSoC6

[KoYa_1937921]

1. Summary

This demo

- inputs images in CIFAR-10 test image dataset saved in Flash of PSoC62,

- classifies the input images into 10 classes with Convolutional Neaural Network (CNN), and

- outputs a result of the classification to LCD display.

2. Hardware required

- Cypress PSoC6 WiFi-BT Pioneer Kit CY8CKIT-062-WiFi-BT

- Cypress TFT Display Shield Board CY8CKIT-028-TFT

3. Software required

- Arm MDK V5.25.2.0, and Cypress PSoC Creator 4.2

- Arm NN Example CIFAR-10 of Arm MDK ( downloadable from Arm GitHub too )

- CE222221 Voice Recorder Example of Cypress PSoC Creator 4.2

Please see the attachment for details:

- AN-PSoC62ArmNNExampleCIFAR10PortingEng.pdf --- Overview and Porting Procedure

- AN-PSoC62ArmNNExampleCIFAR10.mp4 --- Demo Video 1 with CapSense Manipulation

- AN-PSoC62ArmNNExampleCIFAR10r1.mp4 --- Demo Video 2 focused to Display

- ArmCM4FPCnnCifar10r1Lib.zip --- a MDK project file for CNN

- CE222221pCnnCifar10Graphics --- a PSoC Creator project of this demo

[Charles_Lai]

Hi,

Thanks for sharing!

It's impressive implementing CNN classification on PSoC 6 and using it on CIFAR-10 dataset.

It's a pretty good demo illustrating how to integrate Arm MDK to PSoC 6.

Wish to see more explorations and innovations of these coming!

<<<<<<<<<<<<<>>>>>>>>>>>>>

Sincere regards from​ C. L.

<<<<<<<<<<<<<>>>>>>>>>>>>>