imxieyi / waifu2x-ios

iOS Core ML implementation of waifu2x
MIT License
551 stars 58 forks source link

output result is noisy.. #13

Closed farazBhatti closed 4 years ago

farazBhatti commented 4 years ago

Hi, iver tried running your converted coreml models but i am not getting desired results. input image (100*100) Webp net-resizeimage

output image (200*200)

oie_jGgceZazR65k

imxieyi commented 4 years ago

This looks like you are using wrong image channel format. Please refer to implementation in this repo.

farazBhatti commented 4 years ago

@imxieyi , yes ive made sure that image as input to model is of RGB format but still the output is the same.

imxieyi commented 4 years ago

Please refer to documentation on Core ML. Apparently image data you are feeding to the model is in wrong format. (This model needs RRRGGGBBB not RGBRGBRGB format)