标题生成 封面生成
comparing different image caption models in which you have a bunch of models ready to use
template extraction, neural template generation
封面来源：
利用标题进行图片搜索 其实只能站内搜索 因为站外没有这种图片与文字的对应关系
截取视频截图
b站原图 histogram match 20% 去掉文字 镜像反转 加入随机噪声 旋转1度
利用封面进行图片反向搜索 效果其实不好 并没有想要的照片 只能找到原图 有可能起到去水印的效果 但是有限
reverse image search engine
telegram reverse image search bot
neural template gen is a natural language generator based on templates from harvard nlp, can be used for title generation
根据标签生成广告 同样可以根据标签生成视频标题（推荐） 在千言数据集上训练过
https://huggingface.co/cocoshe/gpt2-chinese-gen-ads-by-keywords?text=My+name+is+Clara+and+I+am
title generator(from description):
https://github.com/harveyaot/DianJing/blob/master/scripts/title_generation_lm.py
https://blog.csdn.net/stay_foolish12/article/details/111661358
cover generation
rectangle packing allow overlapping
when solution is not found, decrease the size of rectangles.
youtube title generator using AI:
https://github.com/gdemos01/YoutubeVideoIdeasGeneratorAI
ai thumbnail generator using pyscenedetect:
https://github.com/yoonhero/ai-thumbnail-generator
image captioning:
https://github.com/ruotianluo/ImageCaptioning.pytorch
youzan clip product title generation:
https://huggingface.co/youzanai/clip-product-title-chinese
paper title generator without description:
https://github.com/csinva/gpt2-paper-title-generator
image captioning using cnn and rnn:
https://github.com/SCK22/image_and_video
image captioning can also be used for video captioning. but that will suffice the accuracy.
keras.io image captioning
https://keras.io/examples/vision/image_captioning/
generate image captions using CLIP and GPT(on medium, click continue reading)
https://towardsai.net/p/l/image-captioning-with-clip-and-gpt
gpt3demo.com has provided a lot of interesting tasks that gpt3 can do. including image captioning. may find video captioning, video classification.
gpt3demo.com provided image captioning libs:
https://gpt3demo.com/category/image-captioning
clipclap
gpt-3 x image captions
visualgpt: generate image captions
https://github.com/Vision-CAIR/VisualGPT
generate stories from pictures, using image transformers and gpt-2, just intro no code