image.md

Best Image Format - JPEG-XL

• Converter https://github.com/google/brunsli
• File sizes
  • https://www.reddit.com/r/jpegxl/comments/fjddrr/lossless_image_formats_comparison_webp_jpeg_xl/
  • https://docs.google.com/spreadsheets/d/1ju4q1WkaXT7WoxZINmQpf4ElgMD2VMlqeDN2DuZ6yJ8/edit
• Base formats
  • FLIF and https://github.com/cloudinary/fuif
  • WebP and https://github.com/google/pik

Hypothetical Machine Learning Replacement

• https://github.com/caoscott/SReC
• https://github.com/fab-jul/L3C-PyTorch

The best Perceptive Hash Tool

• https://github.com/thorn-oss/perception
• reasoning: pHash, aHash, dHash and wHash are too common

The best Image Augmentation Tools

• https://github.com/aleju/imgaug
• https://github.com/albumentations-team/albumentations
• https://github.com/mdbloice/Augmentor
• these three all have at least 500 stars

The best Steganography Tools

• https://github.com/DominicBreuker/stego-toolkit
• the best collection out there

The best Neural Denoising Tools

• some info
  • https://github.com/wenbihan/reproducible-image-denoising-state-of-the-art
  • https://github.com/z-bingo/awesome-image-denoising-state-of-the-art
  • https://github.com/flyywh/Image-Denoising-State-of-the-art
  • https://github.com/cszn/KAIR
• Super fast by intel https://github.com/OpenImageDenoise/oidn
• https://github.com/yu4u/noise2noise and https://arxiv.org/pdf/1803.04189.pdf
• https://github.com/GuoShi28/CBDNet and https://arxiv.org/pdf/1807.04686
• https://github.com/wbhu/DnCNN-tensorflow and https://arxiv.org/pdf/1608.03981
• https://github.com/cszn/FFDNet and https://arxiv.org/pdf/1710.04026.pdf
• https://github.com/google/burst-denoising
• https://github.com/manumathewthomas/ImageDenoisingGAN
• https://github.com/rajarsheem/libsdae-autoencoder-tensorflow
• https://github.com/hwalsuklee/tensorflow-mnist-CVAE and https://arxiv.org/pdf/1406.5298.pdf
• https://github.com/czbiohub/noise2self and https://arxiv.org/pdf/1901.11365.pdf
• https://github.com/NVlabs/selfsupervised-denoising and https://arxiv.org/pdf/1901.10277.pdf
• https://github.com/hejingwenhejingwen/AdaFM
• https://github.com/wblgers/tensorflow_stacked_denoising_autoencoder
• https://github.com/Oracen/MIDAS
• https://github.com/guochengqian/TENet and https://arxiv.org/pdf/1905.02538.pdf
• https://github.com/gidariss/wDAE_GNN_FewShot and https://arxiv.org/pdf/1905.01102.pdf
• https://github.com/mgharbi/demosaicnet_caffe
• https://github.com/saeed-anwar/RIDNet and https://arxiv.org/pdf/1904.07396.pdf
• https://github.com/hellloxiaotian/BRDNet
• https://github.com/yzhouas/PD-Denoising-pytorch

The best Neural Image Restoration

• TBD

The best Neural Image Architectures

• https://github.com/BertMoons/Comparing-CNN-Architectures
• https://github.com/CeLuigi/models-comparison.pytorch
• https://github.com/yeephycho/nasnet-tensorflow
• (Did not browse through GitHub for this... papers instead)

The best Semantic Segmentation

• https://github.com/mrgloom/awesome-semantic-segmentation
• pre-cursor https://github.com/amusi/awesome-object-detection
• and also https://github.com/hoya012/deep_learning_object_detection
• also do some reading on "instance segmentation" and "region proposal"
• example repo https://github.com/facebookresearch/detectron2

The best Image Annotators

• https://github.com/tzutalin/labelImg
• https://github.com/wkentaro/labelme

The best Super-resolution Tool

• active test https://github.com/icpm/super-resolution
• https://github.com/YapengTian/Single-Image-Super-Resolution
• https://github.com/huangzehao/Super-Resolution.Benckmark
• https://github.com/ChaofWang/Awesome-Super-Resolution

The best image region tagging

• Source of inspiration https://github.com/facebookresearch/detectron2
• Tool https://github.com/KichangKim/DeepDanbooru
• Visualization https://github.com/halcy/DeepDanbooruActivationMaps

The best image adversarial tools

• https://github.com/tensorflow/cleverhans
• https://github.com/bethgelab/foolbox
• https://github.com/advboxes/AdvBox
• https://github.com/shubhomoydas/ad_examples
• reading list
  • https://github.com/chawins/Adversarial-Examples-Reading-List
  • https://github.com/chbrian/awesome-adversarial-examples-dl
• why avoiding it is impossible https://github.com/anishathalye/obfuscated-gradients

The best NSFW detection tool

• dataset https://github.com/alex000kim/nsfw_data_scraper
  • 2nd place https://github.com/EBazarov/nsfw_data_source_urls
• interface https://github.com/infinitered/nsfwjs
  • Model https://github.com/GantMan/nsfw_model
  • Model https://github.com/rockyzhengwu/nsfw
• Bad version https://github.com/yahoo/open_nsfw
  • Modernized https://github.com/mdietrichstein/tensorflow-open_nsfw

The best forbidden obfuscation software - GMask

• C++ https://github.com/kunstmusik/cmask
• Go https://github.com/fggp/gmask
• Python https://github.com/GunioRobot/pymask
• JS https://github.com/julianwachholz/wmask

Photo unedit

• TBH

Youtube Poop Techniques

• TBH

The best face reference

• https://github.com/ChanChiChoi/awesome-Face_Recognition
• https://github.com/betars/Face-Resources
• https://github.com/becauseofAI/HelloFace
• https://github.com/polarisZhao/awesome-face
• https://github.com/mrgloom/Face-Swap
• https://github.com/mrgloom/Face-landmarks-detection-benchmark
• https://github.com/ShownX/FacePaperCollection
• https://github.com/L706077/DNN-Face-Recognition-Papers

The best 3d facial landmark / face alignment

• https://github.com/deepinsight/insightface
• https://github.com/YadiraF/PRNet
• https://github.com/1adrianb/face-alignment
• https://github.com/cleardusk/3DDFA

The best adversarial face attack

• https://github.com/ppwwyyxx/Adversarial-Face-Attack
• https://github.com/papermsucode/advhat

A good viewer

• https://github.com/exzhawk/EhViewer

The future of image browsing

• Personal in Python https://github.com/hydrusnetwork/hydrus
• Group in Elixir https://github.com/derpibooru/philomena
• Python-Elixir bridge https://github.com/Pyrlang/Pyrlang

The largest dataset

• ImageNet has ~14M http://image-net.org/about-stats
• Places has ~10M http://places2.csail.mit.edu/
• OpenImages has ~9M https://storage.googleapis.com/openimages/web/factsfigures.html
• Danbooru has ~4M https://www.gwern.net/Danbooru2019
• Objects365 has ~2M https://www.objects365.org/overview.html

Discord scrape

• https://github.com/Tyrrrz/DiscordChatExporter

Hentai Cleaning

• https://github.com/natethegreate/Screentone-Remover
• https://github.com/natethegreate/hent-AI
• https://github.com/deeppomf/DeepCreamPy