how to classify imageData Store images depending on each image's object shape
Ältere Kommentare anzeigen
I have ImageDataStore that has 1000 images and looking to classify each images depending on the shape of the object inside each image instead of classifying depending on RCS
Antworten (1)
Image Analyst
am 22 Apr. 2021
0 Stimmen
See my attached shape recognition demo. If you need more help, attach your images and say what shape(s) you want to find in them, AFTER you read this:
6 Kommentare
Mohamed Elbeialy
am 22 Apr. 2021
Mohamed Elbeialy
am 22 Apr. 2021
Image Analyst
am 22 Apr. 2021
I'd probably threshold to get the flower and get a mask of the largest blob. Determine the thresholds by analyzing the colors around the edge of the image and in the central part and finding the color half way between. OOr ir all your flowers are red, just take the red channel and call imbinarize, followed by imfill() and bwareafilt() to get the largest blob. Then I'd get the boundaries with bwboundaries() and find the disances of each boundary point ot the centroid. Do this for each flower to make up a reference library of distance signals. Then I'd use normalized cross correlation with the distance signal from an unknown flower to see which reference flower it matches up best with.
Mohamed Elbeialy
am 22 Apr. 2021
Image Analyst
am 23 Apr. 2021
Not sure I'll have time. What's your "use case"? Why do you need to do this? And where/how are you getting the photos.
Mohamed Elbeialy
am 23 Apr. 2021
Kategorien
Mehr zu Deep Learning Toolbox finden Sie in Hilfe-Center und File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!



