[{"data":1,"prerenderedAt":307},["ShallowReactive",2],{"project-26118756":3},{"id":4,"name":5,"line":6,"line_source":7,"title":8,"url":9,"domain":10,"maker":11,"posted_at":12,"year":13,"points":14,"comments":15,"rank_in_year":16,"shown_in_year":17,"top":18,"tier":19,"category":20,"subcategory":21,"format":22,"keywords":23,"reception":28,"sentiment":29,"stance_mix":30,"scored_comments":34,"aspects":35,"strip":42,"post_text":44,"what":45,"said":73,"maker_in_thread":156,"shape":167,"grid":170,"attempts":267,"same_week":268,"maker_shows":37,"maker_hidden":299,"reader_version":300,"card":301},"26118756","natural-language-youtube-search","Search inside YouTube videos using natural language queries","title","Show HN: Search inside YouTube videos using natural language queries","https:\u002F\u002Fgithub.com\u002Fhaltakov\u002Fnatural-language-youtube-search","github.com","vladoh",1613165342,2021,264,43,140,2533,6,"legend","ai","image and video search","unknown",[24,25,26,27],"clip","video search","natural language search","youtube","curious",0.034,{"positive":18,"neutral":31,"mixed":32,"critical":33},10,0,3,19,{"design":36,"execution":38,"performance":39,"idea":40},{"positive":37,"critical":32},1,{"positive":32,"critical":37},{"positive":37,"critical":32},{"positive":41,"critical":37},4,[32,33,32,32,43,33,32,33,32,33,33,32,32,32,43,33,33,43,32,43,32,33,32,43,32,33,32,43,32,32,32,37,32,32,32,32,37,33,32,37],5,null,{"problem":46,"audience":44,"pricing":44,"motivation":53,"time_to_build":44,"stack":57,"openness":22,"stage":72,"team_size":22},{"value":47,"source":48,"evidence":49},"Finding specific frames inside a YouTube video using natural language queries","maker",[50],{"id":51,"by":11,"text":52},"26118968","This is a demonstration of the abilities of OpenAI's CLIP neural network. The tool will download a YouTube video, extract frames at regular intervals and precompute the feature…",{"value":54,"source":48,"evidence":55},"Demonstrate the abilities of OpenAI's CLIP neural network",[56],{"id":51,"by":11,"text":52},[58,62,68],{"value":59,"source":48,"evidence":60},"OpenAI CLIP",[61],{"id":51,"by":11,"text":52},{"value":63,"source":48,"evidence":64},"PyTorch 1.7.1",[65],{"id":66,"by":11,"text":67},"26119701","Yes, I know that this is a bit slow. The problem is you really need 1.7.1, because 1.7.0 leads to some strange issues and broken results:…",{"value":69,"source":48,"evidence":70},"Google Colab notebook",[71],{"id":51,"by":11,"text":52},"prototype",{"praise":74,"objections":95,"requests":113,"questions":135,"alternatives":149},[75,88],{"value":76,"source":77,"evidence":78},"Impressive results of CLIP-based frame search","commenter",[79,80,84],{"id":51,"by":11,"text":52},{"id":81,"by":82,"text":83},"26120255","mockingbirdy","Awesome! I'm currently working on the exact same thing (but with OCR added). Thank you for releasing this.",{"id":85,"by":86,"text":87},"26121528","ramraj07","This is amazing. I'm going to get this running on my Dropbox. Curious what it gets out of scanned documents as well.",{"value":89,"source":77,"evidence":90},"Heatmap of query scores per frame adds value",[91],{"id":92,"by":93,"text":94},"26172115","pininja","Looks great! Adds a whole dimension to the video. Thanks.",[96,106],{"value":97,"source":77,"evidence":98},"Downloading YouTube videos may violate YouTube's terms of use",[99,103],{"id":100,"by":101,"text":102},"26122594","gbennett71","Downloading YouTube videos is a Terms of Use violation, no?",{"id":104,"by":101,"text":105},"26130445","https:\u002F\u002Fwww.youtube.com\u002Fstatic?template=terms&gl=AU Permissions and Restrictions You may access and use the Service as made available to you, as long as you comply with this…",{"value":107,"source":77,"evidence":108},"Scanning many videos is slow and does not scale",[109],{"id":110,"by":111,"text":112},"26120029","mandeepj","> The tool will download a YouTube video, extract frames at regular intervals That should be able to scale well :-)",[114,121,129],{"value":115,"source":77,"evidence":116},"Search across all of YouTube, not just a provided video",[117],{"id":118,"by":119,"text":120},"26120944","suyash","The problem still exists that you have to provide it the YouTube video to search within, would be nice if there was a tool to search across all of YouTube.",{"value":122,"source":77,"evidence":123},"Support for OCR text search in video",[124,125],{"id":81,"by":82,"text":83},{"id":126,"by":127,"text":128},"26120741","canada_dry","> the exact same thing (but with OCR added) Hmmm... what does \"with OCR added\" mean? If there is text in the video (e.g. street sign) that it can also be searched??",{"value":130,"source":77,"evidence":131},"Smarter frame extraction such as keyframes",[132],{"id":133,"by":127,"text":134},"26120766","Great demo. Wondering whether it would be more efficient if extracting frames where the content has changed (e.g. over a threshold and\u002For all I-frames)? Also, could this be used…",[136,142],{"value":137,"source":77,"evidence":138},"Does it produce an occurrence index or word map",[139],{"id":140,"by":93,"text":141},"26119047","This is very cool! Does this produce a occurrence index by any chance? It would be neat to explore a word map of a video.",{"value":143,"source":77,"evidence":144},"Can it detect complex events like traffic violations",[145],{"id":146,"by":147,"text":148},"26121785","Crazyontap","Can it work for more advance keywords like say, \"traffic violation\" where it spots a car jumping red light or pedestrian not using a crosswalk, etc? It could be very useful to…",[150],{"value":151,"source":48,"evidence":152},"Google search linking to YouTube timestamps",[153],{"id":154,"by":11,"text":155},"26126454","CLIP is quite a new model published by OpenAI in January. Their work is novel and pushed the state-of-the-art in this area by a lot. I'm sure that Google is also working on…",{"comments":157,"answered":158,"key":159},17,7,[160,161,164],{"id":51,"text":52},{"id":162,"text":163},"26119072","Not yet, but I had this idea as well. You basically get a score describing how well a phrase is matching each of the images so it won't be difficult to do. I'll look into that!",{"id":165,"text":166},"26119995","I pushed a small update and you will now see a heatmap displaying the score of the search query for each frame.",{"commenters":34,"max_depth":158,"maker_replies":157,"first_comment_mins":168,"lifespan_hours":169},23,120,[171,173,175,176,178,180,184,185,187,190,192,195,197,199,201,204,205,207,209,211,213,214,216,217,219,222,224,227,231,232,236,238,241,243,245,248,249,252,255,257,259,261,265],[51,32,32,172,11,37],"project\u002Fpositive\u002Fidea",[140,32,33,174,93,32],"project\u002Fneutral\u002Fnone",[162,32,32,174,11,37],[165,32,32,177,11,37],"project\u002Fneutral\u002Fdesign",[92,32,43,179,93,32],"project\u002Fpositive\u002Fdesign",[181,32,33,182,183,32],"26119580","project\u002Fneutral\u002Fexecution","fireattack",[66,32,32,182,11,37],[186,32,33,182,183,32],"26119798",[188,32,32,189,11,37],"26119840","commenter\u002Fneutral\u002Fnone",[110,32,33,191,111,32],"project\u002Fneutral\u002Fperformance",[193,32,33,191,194,32],"26122254","woko",[196,32,32,189,11,37],"26122273",[198,32,32,189,194,32],"26122284",[200,32,32,189,11,37],"26123315",[202,32,32,203,194,32],"26124446","other\u002Fneutral\u002Fnone",[81,32,43,172,82,32],[126,32,33,206,127,32],"project\u002Fneutral\u002Fidea",[208,32,33,174,82,32],"26120903",[133,32,43,210,127,32],"project\u002Fpositive\u002Fperformance",[212,32,32,191,11,37],"26122264",[85,32,43,172,86,32],[215,32,32,177,11,37],"26122238",[146,32,33,206,147,32],[218,32,32,206,11,37],"26122232",[220,32,43,172,221,32],"26122623","Abishek_Muthian",[223,32,32,206,11,37],"26123307",[225,32,33,206,226,32],"26127567","akinhwan",[228,37,32,229,230,32],"26124443","other\u002Fneutral\u002Fidea","jameshart",[154,37,32,203,11,37],[233,234,43,172,235,32],"26123536",2,"mncharity",[237,234,32,203,11,37],"26126530",[239,33,32,203,240,32],"26123392","amelius",[242,33,32,189,11,37],"26126479",[100,41,37,244,101,32],"project\u002Fcritical\u002Fnone",[246,41,32,189,247,32],"26123154","hansel_der",[104,41,32,189,101,32],[250,41,32,251,11,37],"26123322","project\u002Fneutral\u002Fprivacy",[253,41,32,189,254,32],"26158396","hddu",[118,43,37,256,119,32],"project\u002Fcritical\u002Fidea",[258,43,33,206,194,32],"26123117",[260,43,32,174,11,37],"26123326",[262,18,37,263,264,32],"26118980","project\u002Fcritical\u002Fexecution","purplecats",[266,18,32,182,11,37],"26119010",[],[269,274,279,284,289,294],{"id":270,"name":271,"points":272,"category":273},"26134915","Khan-dl",473,"learning_science",{"id":275,"name":276,"points":277,"category":278},"26138446","ustaxes.org",465,"web_saas",{"id":280,"name":281,"points":282,"category":283},"26168285","Ray.so",384,"dev_tools",{"id":285,"name":286,"points":287,"category":288},"26154038","Endgame",368,"security_privacy",{"id":290,"name":291,"points":292,"category":293},"26131123","Jam",317,"consumer_social",{"id":295,"name":296,"points":297,"category":298},"26128388","M1 Chart",205,"finance_crypto",false,"v2",{"id":302,"name":5,"line":6,"ai":299,"cat":20,"year":13,"points":14,"comments":15,"rec":28,"s":303,"top":18,"kw":304,"maker":11,"domain":10,"tier":19,"attempts":32,"strip":305,"url":9,"mix":306,"praise":76,"objection":97,"sub":21},26118756,0.03,[24,25,26,27],"0300530303300053350503050305000100001301",[18,31,32,33],1794036603020]