How to Collect Images for Deep Learning

For Neurocle Deep Learning software, there is no need to be confined to a single brand of camera. The software is flexible enough to work with most image collected.

IMAGE REQUIREMENTS

Format:   JPG(JPEG), PNG, BMP, TIF(TIFF), DCM

Size:    Smaller than 2048 x 2048 pixels

For this article’s demo, Microscan Gige Camera and Customised Deep Learning UI (Interface) would be used.

Click image to enlarge
CUSTOMISED DEEP LEARNING UI (USER INTERFACE)

Click above image to enlarge
SET THE FILE PATH AND NAME THE FOLDER

Click above image to enlarge
PLACE OBJECT IN CAMERA’S FIELD OF VIEW (FOV). MAKE SURE ITEM OF INTEREST IS CAPTURED.

Click above image to enlarge
CLICK ‘ENABLE SAVE IMAGE’ ONCE READY.

DEMO VIDEO

Click above image to enlarge
CONFIRM YOUR IMAGES IN THE SAVED FILE DIRECTORY
VERIFY THAT THE IMAGES ARE NAMED CORRECTLY.

UPLOADING TO NEURO-T

1. Log in into Neurocle

2. Create Label Set

3. Choose to upload by Images/Folder

4. Image upload complete

Feel free to contact us for more information.

Salesperson: Quek, 9113 3566
Email: sales@zincode.net