SSL92/hyperIQA

Issues

• why test_ROCC and test PLCC descend

  #28 opened 4 years ago by lijun-chiebot
  2

• How to calculate MOS_zscore in koniq10k datasets?

  #15 opened 4 years ago by wangbin2018
  6

• HyperIQA resize

  #48 opened a year ago by zht8506
  0

• about data process

  #47 opened a year ago by DIY-Z
  0

• Can you share the GFIQA dataset?

  #46 opened a year ago by SZP1234
  0

• coda question

  #45 opened 2 years ago by TANGLONG2
  0

• about saving model

  #9 opened 4 years ago by yinglili666
  7

• koniq-10k dataset

  #44 opened 2 years ago by Zhongguangtan
  0

• The rationality of sample patches randomly.

  #43 opened 2 years ago by Cassieyy
  0

• Evaluation of noise image quality

  #42 opened 2 years ago by SummerProgrammer
  0

• MSU Video Quality Metrics Benchmark Invitation

  #41 opened 2 years ago by msm1rnov
  0

• About the random split way of the dataset

  #39 opened 2 years ago by Jasonbins
  0

• About resize in Koniq-10k dataset transforms

  #40 opened 2 years ago by LinpengPan
  0

• GoogleDrive link for pre-trained model

  #36 opened 3 years ago by SheydaGhb
  4

• 关于PLCC的计算

  #38 opened 3 years ago by chencn2020
  0

• Can you share the CSIQ database?

  #37 opened 3 years ago by LinpengPan
  2

• 论文中提到的测试数据集BID在哪下？

  #30 opened 4 years ago by weidelongdongqiang
  4

• Question about the performance of SOTA method.

  #35 opened 3 years ago by xiezhangxiang
  2

• stats.pearsonr Error？

  #34 opened 3 years ago by XiaoqiWang
  1

• How to use one more GPUS

  #26 opened 4 years ago by lijun-chiebot
  2

• .cpu可以改成.cuda吗

  #33 opened 4 years ago by sng6238
  0

• HyperNetWork is not being trained

  #32 opened 4 years ago by mleoelm
  2

• trained model

  #13 opened 4 years ago by ashbringerf
  4

• could you please release the training dataset？

  #20 opened 4 years ago by JasonZhangS
  1

• The reproduced result is not consistent with your paper. Is there any problem ?

  #23 opened 4 years ago by yongZheng1723
  1

• Resume training

  #24 opened 4 years ago by pencilzhang
  1

• Sometimes the score will be greater than 100？

  #29 opened 4 years ago by alexandsunny
  1

• Where can I download the BID database?

  #14 opened 4 years ago by lllllllllllll-llll
  1

• Value changes every time

  #31 opened 4 years ago by sujyQ
  1

• train_patch_num

  #27 opened 4 years ago by lijun-chiebot
  0

• 关于demo中resize问题

  #25 opened 4 years ago by alexandsunny
  0

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

  #22 opened 4 years ago by pencilzhang
  2

• 可不可以改成全卷积网络，来适应不同大小图片

  #17 opened 4 years ago by wangbin2018
  12

• In the hyperiqasolver.py file, why Param. Requires_grad = False?

  #16 opened 4 years ago by huahua-c
  4

• How to train model with koniq10k

  #21 opened 4 years ago by xyy19920105
  1

• Can you provide a web address for the KonIQ-10k database? The official link won't download.

  #19 opened 4 years ago by ygmuy
  0

• regarding the logistic regression before plcc

  #18 opened 4 years ago by seyeeet
  3

• how to test an image and get the score?

  #12 opened 4 years ago by yinglili666
  3

• the 'dmos_realigned.mat' of LIVE database

  #11 opened 4 years ago by yinglili666
  5

• where is the "csiq_label.txt"?

  #10 opened 4 years ago by lllllllllllll-llll
  1

• About inference code

  #8 opened 4 years ago by KleinXin
  1

• A question about koniq-10k and bid dataset

  #6 opened 4 years ago by feng150416
  2

• about label/score normalization

  #7 opened 4 years ago by ainnn
  5

• Why setting 'params.requires_grad=False' for target_network?

  #5 opened 4 years ago by luofan18
  1

• Please release the code

  #2 opened 4 years ago by siddhantkatyan
  2

• Problems of Implementation Details. Some difference?

  #4 opened 4 years ago by lllllllllllll-llll
  2

• Push for code~

  #3 opened 4 years ago by sureJ0824
  1

• code?

  #1 opened 4 years ago by upupbo
  10