Why not using 512 x 384 images of koniq10k to avoid resizing?

SSL92 / hyperIQA

Source code for the CVPR'20 paper "Blindly Assess Image Quality in the Wild Guided by A Self-Adaptive Hyper Network"

MIT License

346 stars 52 forks source link

Why not using 512 x 384 images of koniq10k to avoid resizing? #22

Open pencilzhang opened 3 years ago

pencilzhang commented 3 years ago

Hi, thanks for your IQA work! I noticed that you resize the input image of koniq-10k dataset. Since there are two resolutions of data in koniq-10k dataset, if you use 512 x 384 as input size, why not using 512 x 384 image size provided by koniq-10k ?

SSL92 commented 3 years ago

Sure you can use 512x384 images for directly training, it's just because we only downloaded the images of 1024x768 resolution...

pencilzhang commented 3 years ago

@SSL92 ok thanks for you response.