Autonomous Machines & Society.

2022-05-31

Video Delogo_Inpainting

Video delogo/inpainting

propainter remove watermark with tracking, entity segmentation

you can use clip for watermark detection, but you don’t know where. fuck. you better train it yourself.

import timm
# all other pytorch imports
# Not used but necessary when running on images
transforms = T.Compose([
T.Resize((256, 256)),
T.ToTensor(),
T.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
])
# Create model
model = timm.create_model('efficientnet_b3a', pretrained=False, num_classes=2)
model.classifier = nn.Sequential(
nn.Linear(in_features=1536, out_features=625),
nn.ReLU(),
nn.Dropout(p=0.3),
nn.Linear(in_features=625, out_features=256),
nn.ReLU(),
nn.Linear(in_features=256, out_features=2)
)
# Load model weights
state_dict = torch.load('./model.pt')
model.load_state_dict(state_dict).eval().to(device)
# Sample Image
im = torch.randn(8, 3, 256, 256)
with torch.no_grad():
pred = model(im)
syms = F.softmax(pred, dim=1).detach().cpu().numpy().tolist()
for water_sym, clear_sym in syms:
# Do whatever you want with the watermark simlarity

image local contrast enhancement, for removing hard-to-detect watermarks

maybe you can use the same trick (context preserving sliding window) from your search engine to here (image preprocessing)!

paddleocr识别效果最好可以识别水印位置以及文字

Linear Contrast Stretching, HE, AHE, CLAHE of an image using matlab

Histogram Equalization (HE)

Adaptive Histogram Equalization (AHE)

Contrast Limited Adaptive Histogram Equalisation (CLAHE)

experiment path:

pyjom/tests/remove_subtle_watermark_local_contrast_ocr

bing query for image local contrast

darktable lua api and scripting

darktable local contrast darktable is an open-sourced photography postprocessing software

configurations:

details: 443
highlights: 36
shadows: 25
midtone range: 0.16

imagej clahe local contrast enhancement

l2uwe L^2UWE: A Framework for the Efficient Enhancement of Low-Light Underwater Images Using Local Contrast and Multi-Scale Fusion written in matlab

glcae Global and Local Contrast Adaptive Enhancement for Non-uniform Illumination Color Images in python

CNN-Based-X-ray-Morphological-Decomposition

github query for local image contrast

image processing basics Image Reading, writing, histogram, histogram equalization, local histogram equalization, low pass filter, high pass filter, geometrical transformation

contrast normalization is an implementation that applies local contrast normalization to images in matlab

contrast enhancement as a github topic

mclahe NumPy and Tensorflow implementation of the Multidimensional Contrast Limited Adaptive Histogram Equalization (MCLAHE) procedure

deepcontrast A deep learning-based fully-automatic intravenous contrast detection tool for head-and-neck and chest CT scans.

mirnetv2 (TPAMI 2022) Learning Enriched Features for Fast Image Restoration and Enhancement. Results on Defocus Deblurring, Denoising, Super-resolution, and image enhancement

pymusica is a contrast enhancement approach involving non linear mapping of Laplacian pyramid.

imWeightedThresholdedheq attempts to enhance contrast of a given image or video by employing a method called weighted thresholded histogram equalization (WTHE).

imagemagick wand local_contrast function

dual gamma clahe Automatic Contrast-Limited Adaptive Histogram Equalization With Dual Gamma Correction

imhblpce attempts to enhance contrast of a given image by employing a method called HBLPCE.

matlab localcontrast for image

global contrast enhancement

im2dhiseq attempts to enhance contrast of a given image by equalizing its two dimensional histogram.

previous research

deeplearning_inpainting:

https://github.com/Sanster/lama-cleaner

ffmpeg delogo:

https://www.jianshu.com/p/2eb1811b5fc6

https://hhsprings.bitbucket.io/docs/programming/examples/ffmpeg/blurring_unsharping/delogo_removelogo.html

https://securitronlinux.com/debian-testing/remove-a-logo-from-a-video-easily-with-ffmpeg/

opencv inpainting/blurring with edge blending

opencv morphlogical operations:

https://pyimagesearch.com/2021/04/28/opencv-morphological-operations/

2022-05-31

Algorithms Compilers Sicp Clrs Ct4S

the algorithms

introduction to algorithms clrs 4th edition in python:

https://zhuanlan.zhihu.com/p/466819939

official clrs 4th edition python code

clrs 4th pdf

clrs 3th edition python code:

https://github.com/tonywangcn/Introduction-to-Algorithms-3rd-Edition-python-code

sicp in python gitbooks:

https://wizardforcel.gitbooks.io/sicp-in-python/content/18.html

algorithms 4th edition official java unofficial python code:

https://github.com/kevin-wayne/algs4

https://github.com/shellfly/algs4-py

https://github.com/itu-algorithms/itu.algs4

2022-05-31

Typemonkey 字说 Oss Alternative

javascript version in web:

https://nostarsnow.github.io/2019/01/20/typemonkey/

https://github.com/nostarsnow/typeMonkey.js

2022-05-31

Optical Flow

flownet nvidia

nvidia optical flow sdk supports all turing gpus (like gtx1660) and above except for gtx1650(tu117).

mmflow from openmmlab:

https://mmflow.readthedocs.io/en/latest/

全自动电影解说软件介绍全自动短视频合成

自媒体自学网新媒体运营自动化运营混剪：

https://www.zmtzxw.com

generate video from plain text using beatdetectorforgames:

https://github.com/FireFragment/video-generator

youtube reddit text to speech video generator:

https://github.com/HA6Bots/Automatic-Youtube-Reddit-Text-To-Speech-Video-Generator-and-Uploader

twitch clip compilation:

https://github.com/HA6Bots/Twitch-Clips-Compilation-Generator-TCCG-

slideshow video generator:

https://github.com/oknoorap/vidshow

best of twitch video generator:

https://github.com/BayoDev/Twitch-Best-Of-Gen

video generator by scraping reddit videos and comments:

https://github.com/charlypoirier/redditube

youtube video generation based on watson natural languahe understanding and google image search:

https://github.com/rhenriquea/ai-video-generator

tiktok video compilation based on custom filters:

https://github.com/HA6Bots/TikTok-Compilation-Video-Generator

12个搬运手法：

去水印放大缩小镜像抽帧加滤镜加特效调色调速转场调整视频顺序加画中画其他视频元素

文案加字幕朗读方案：

https://m.baidu.com/video/page?pd=video_page&nid=9603074179371472094&sign=5664314656417860263&word=自动合成视频&oword=自动合成视频&atn=index&frsrcid=4185&ext=%7B"jsy"%3A1%7D&top=%7B"sfhs"%3A1%2C"_hold"%3A2%7D&sl=4&fr0=A&fr1=A&ms=1&lid=12061699393737547668&referlid=12061699393737547668&frorder=6&_t=1653970095475

知乎方案，模板加图片，BGM，配音

文章自动生成视频，幕言自动打轴（force alignment）：

https://zhuanlan.zhihu.com/p/218000255

https://www.muyanpeiyin.com/?zhihu

5种方式生成原创视频：

https://zhuanlan.zhihu.com/p/140075360?from_voters_page=true

https://zhuanlan.zhihu.com/p/119422109

https://m.baidu.com/video/page?pd=video_page&nid=11383613456893718608&sign=855254448776210471&word=AI全自动剪辑软件：2分钟自动合成一个影视解说视频，批量效率高&oword=自动合成视频&atn=index&frsrcid=5373&ext=%7B"jsy"%3A1%7D&top=%7B"sfhs"%3A1%2C"_hold"%3A2%7D&sl=4&fr0=A&fr1=C&title=AI全自动剪辑软件：2分钟自动合成一个影视解说视频，批量效率高&lid=12061699393737547668&ms=1&_t=1653971026899

模式一：单视频+文案

一个视频配一个文案，软件将自动把文案转化成字幕以及真人发音，把视频和字幕添加到视频上

此模式适合电影解说，新闻讲解等等

模式二：多视频+文案

多个视频会随机合成一个视频,然后把文案转化成字幕以及真人发音，把视频和字幕添加到视频上

此模式适合抖音带货等

模式三：多图片+文案

多个图片会随机合成一个视频,然后把文案转化成字幕以及真人发音，把视频和字幕添加到视频上

此模式适合大批量却又苦于找不到视频素材的情况

模式四：单视频混剪

适合抖音带货，或者搬运类的影视混剪

模式五：文字转语音

现在头条也出了音频的创作专区，应广大用户的需求，添加这个功能，将txt文件批量放入video目录即可。

2022-05-31

Deepnude Censorship Nsfw Safesearch

clip-based-nsfw-detector

you can turn on/off safesearch by changing ip to iran. is there better options, doing this programmatically?

nsfw createml imageclassifier with pretrained model

敏感词过滤器 also need a high speed censor engine using trie, also in diffferent language(like english or japanese)

english profanity:

https://github.com/snguyenthanh/better_profanity

https://github.com/zacanger/profane-words

https://github.com/MauriceButler/badwords

https://github.com/vzhou842/profanity-check

https://github.com/coffee-and-fun/google-profanity-words

chinese profanity:

https://github.com/gaohuifeng/sensitive-word-filter

https://github.com/lunzima/profanities.txt

https://github.com/nyx1987/forbiddenwords

https://github.com/insoxin/bannedwords

https://github.com/chason777777/mgck

https://github.com/k5h9999/keywordfilter

https://github.com/tomzhang/bannedwords

https://github.com/observerss/textfilter （需要回看历史查看git历史）

https://github.com/aojiaotage/text-censor

deepnude nsfw nude picture detection:

https://github.com/yuanxiaosc/DeepNude-an-Image-to-Image-technology/blob/master/README-ZH.md

videoaudit 视频审核框架https://github.com/minitrill/VideoAudit

nsfw: not safe for work, inappropriate content, porn, offensive

https://github.com/rockyzhengwu/nsfw

https://github.com/alex000kim/nsfw_data_scraper

https://github.com/yahoo/open_nsfw

https://github.com/Rayraegah/nsfw_japan

https://github.com/fishsup/nsfw-image-classification

https://github.com/yangbisheng2009/nsfw-resnet

https://github.com/GantMan/nsfw_model

https://github.com/devzwy/open_nsfw_android

https://github.com/infinitered/nsfwjs

https://github.com/nsfw-filter/nsfw-filter

nudity, violence and drug

https://github.com/amshrbo/nsfw-detection

to train these networks, suitable datasets are required.

找训练集找涉政涉黄暴力血腥图片训练集找类似的文字训练集百度aistudio可能有 github可能有百度一下也可能有

同样的思路情绪情感打分也可以这样打分根据不同的训练集进行打分

github violence detection:

https://github.com/topics/violence-detection

Bloody Image Classification with Global and Local Features

https://www.researchgate.net/publication/309365631_Bloody_Image_Classification_with_Global_and_Local_Features

Object content understanding in images and videos draws more and more attention nowadays. However, only few existing methods have addressed the problem of bloody scene detection in images. Along with the widespread popularity of the Internet, violent contents have affected our daily life. In this paper, we propose region-based techniques to identify a color image being bloody or not. Firstly, we have established a new dataset containing 25431 bloody images and 25431 non-bloody images. These annotated images are derived from the Violent Scenes Dataset, a public shared dataset for violent scenes detection in Hollywood movies and web videos. Secondly, we design a bloody image classification method with global visual features using Support Vector Machines. Thirdly, we also construct a novel bloody region identification approach using Convolutional Neural Networks. Finally, comparative experiments show that bloody image classification with local features is more effective.

search for nsfw filter providers on google/kaggle

kaggle nsfw image dataset

https://www.kaggle.com/datasets/laxmansingh/nsfw-images-data

https://www.kaggle.com/datasets/drakedtrex/my-nsfw-dataset/code

search for nsfw image/text on github:

https://github.com/alex000kim/nsfw_data_scraper

https://github.com/nsfw-filter/nsfw-filter

nsfwjs

https://github.com/arufian/Image-Censor-Lightning-Web-Component

https://github.com/enymuss/censorText

https://github.com/fmsky/resnet50_inappropriate_content_detect

https://github.com/CheranMahalingam/Image_Content_Moderation