📜 ⬆️ ⬇️

How many cats on Habré?

Recently, I was traveling by bus from Toronto to New York, it was dark outside, there was a bit of port inside, I didn’t want to sleep at all, and I decided to sort out Deep Learning. I downloaded Caffe , fed him a couple of pictures, on which the ball and banana were correctly recognized. I wanted to recognize something more interesting, and I remembered that somewhere on the hard drive I had a dump of a habrahabr, which I did when I was in the course of information retrieval in the ShAD Yandex.

It took several minutes to write a script that recognizes what is depicted on the avatar of a habial user and felines all cats, and it took several days to update the dump to the current one and the recognition of images, and now I can say that there are at least 748 cats in it.

Under Habrokat you can read a little more details and look at all the cats.
')




Within a few days, my spider received a habrahabr, and when I saw that I had over 500,000 pages of habra users downloaded, I stopped it. As a result, I had information about 527468 users. Of these, 73193 were avatars.

Next, with the help of Caffe, I received for each picture the first five most likely tags. After filtering by cat tags, 1013 avatars remained. Among them were some numbers of non-cats, for example, owls are often recognized as cats.


Alf, no matter how funny, half primate, and half a cat. Classes are defined as: 'primate', 'monkey', 'lion', 'Old World monkey', 'feline'.


But Zelenyikot is not a cat, the closest that is recognized is the 'giant panda'.


After I manually removed erroneously credited cats, 748 avatars remained. All of them are shown in the picture below.



List of cats
1cc alexsunsky , alexvarp , alienattack , alkair , alsmile , alvast , alx_voronov , ankhema , anaphema , anaphema , anaphema , andyapanda82 , annael , anesthero artemchik , arthurstark , artic , artyomcool , asaks , askadias , asker1987 , assc , astass , at0m1c , atpmodern , @ave, aver_pasha , avlaak , awwes , azazel17 , badcentlemen , bad_student , bagas , bar , zakel17 , bar , bar , bar , bar , bar , bar , bar , bar , bar , bar , bar , bar , bar , bar , bar , jaw , bazeyvit , bazlan , bearer , benign , bennyfinch , bereg , beretta , bescom , betepoknoname , bioluh , biorobot , bizdirect , blayder , blazer44 , blondinekatze , bloomberg_ door , boban1951 , boeing777 , booldogg , borland34 , bristly , bsrginru , bushuevrussia , butylski , byabuzyak , byv81 , catmen08 , cat_of_castle , catplusplus , c-darwin , cdman2010 , celldweller24 , ch3sh1 , ky , chy , csh darwin , cdmanwin0 ckuridin comargo compoz comrade_55 contango controlflow corpix ded_mopo3 , deeman313 , deeptds , deepx , deleted-krakazyabl , demalah , demosfobos , dempil , denis__sh , denv , denvo , denzik , deorbit , derol , desher , dgad , digvice , di_halt , dikah , dikeyfd , dimf , dhf , dhf , dgad , digik divinity32 , d_landau , dmitrychast , dmitryga , doberbeatz , dolphinsoft , drgosh_m , droll13 , druh , dsk1306 , dyumba , eaterman99 , ebashu_na_bayane , echoes11 , ed-fruty , eg iptyanin , egorikua , ejz , ekup , elektrochajnik , elfets , elnee , turks , erkowa , erud , eugenco , eugeniyazhenka , eugenyh , eujin , evolutin , exl , extender , extra300 , f1ex , f7soft , dept- apt , e , jolin , evolutin , eludin , erudin fkysn9wa , flowergirl , foka , folond , fomkadev , forestcat , forever , forgettmenot , foxus , freelock , frikazoid , frumarina , gaparchi , gazizov , gd666 , gekty , gena_tcvali ; gkovaleff , glazunov_v , gliese581 , goldkanet , goldkeeper , goodwish , gorbln , grabse , graf_duku , grammidin , gramoboi39 , greencat , gregory_epstein , gregosh , gumja , guyvernk , gwee , habralan , habrwaterman , my , my , my , my , mygosh , gumja , hagralan hoffman , hollyjohn , homphsexception , hrakk , ibice , icefroz , idimka39 , idolopoklonnica , idtm , ig_gor , ihard , iiii , ilirio , iluhinhabr , ilyapenkov , i-marketing , im_nowhare , inbider , i ndivid1 , ingift , inikulin , innominate , ioscat , irbisfx , iser , it_fake , itt2 , ivachkin , ivanmarkov , ivasyai , ivinnic , iwanerhov , jabberwock , jshinj , jaguar6868 , jaster_ru , jazzpixels , jerseyzx , j95 j , pix j joossy julsky k1b0rg klaizar , klazfi kultishev , kuprikow , kuzia9 , kvil , kvo , kwachboba , kww , ky4web , kyver , la6puk , lakostik , lamkaslim , lanaluna , laroux , lastromantik , lavash , ldvcl , le0nidy4 , leeroy , lefechki ; ey , lemuil , lemurija , lena_sakhno , leo_sgm , lexinc , lija , likenoother , liquidstorm , lirein , loginsin , logofiles , lonelycat , lonely_tramp , longedok , longnull , lova , lucha , luciferr , luckycoin , lukrets ,,, and , longnull , lova , lucha , luciferr , lichlyr , lys , petrol mackor , madfortune , madgnu , madhacker , madmaximka , madmit , mahowik , makahsahc , maks78 , maksimau , mak_vit , manaxname , manulable , manul , mardy_bum , margin0px , maria_platonova , markstan , mattheus , mauschka , makeh , a, hake , makeh , we , 2014 , xy megamonster99 , megaromb , menlod , meph1st0 , metzgermeister , mezhevikin , mi4log , mikle , minmatarec , mintonist , mitcenika , mkot , moll , molodoy_ua , mr_gruber , mrjovas , mrzdgg , muhammad_97 , my businesses , my businesses , my businesses , my businesses . my_username , nardin7 , nardin , nart09 , nastika_namaste , natali_khatkevich , natfomm , ne_kotin , neozoid , neronmoon , nickvenator , nightlion , nik684 , nikiomen , nik-n ikita , nklnke , nlait , nobr , nomadmoon , norgen , noskivpolo phantasmx quartz , r2ot , radiokot , radioniker , rasifiel , rasswet , rastaropnaya , rayvor , razvor , real_voland , rebulus , recosh , refakki , reflash , remuktus , respown , retuned , rez1dent3 , rezwyi , rhorobets , ie , i will apply , reeu romanitalian_net , romanvl , romanvz , romens , root1564 , rour , rpex , rsilaev , rtaburceanu , rubik , rusakovnikita , ruslanm , rv630 , s1lent_ua , sabiovradik , sad_robot , sagibshikar , Sagifire , sammily , san41to , sanchez17_08 , sandman , satellence , saturday_cinderella , saul , sav13 , savann , savvy , sax , scanderberg , schigartsov , schrodingers_cat , scorp2011 , sealse , secruity , seeker142 , semisay , sempron , sergdrakon , sergeich65 , serge_k , sergeyam sergei , we are not in love , sid44 , siddhartha , simple_rb , sirleh , skif300000 , skydreamer , smkuzmin , snowleopardik , snuk182 , sob_s , sokovenko , solteks , somikosa , sonkelf , sonygirl ; , stasabrosimov , stdio , stemrus , stixnematic , stjuart , strannik7j , superhabrik , surendil , svd , svetlanaang , svetlyiakapro , svinoki tiger tigrasan , tigrik , timiriliyev , timofffey , timurnn , tirael78 vampire666 , vanatka , vasi , vidiriv , vidiriv , vdv , vdv , vdv , vdv , vdv , vdv , vdv , vdv , vdv , vdv , vdv , vdv , vdv , vdv , vdv , vdv , vdc wadman , wadyan0i , wantbefree , wase4ka , webcote , welcomeroot , wentswc , whitetigr , wildraid , witjaz007 , wmas , wohlstand , wonderkot , x3n , xakka , xakpc , xa lasdor , xanderslex , xavadonx , x__cross , xvadim , yadowit , yami513 , yanis , yaroslav_dbls , yaruson , yazvin , yurapb , yurc , yushkin , zaynetro , zelik_bog , zellien , zentif , zs000 , zshchye , zurtif , zelik_bog , zellien , zentif , zsik , zshnet , zyenro , zeltif , zeltif , zeltif , zentif , zushar

In conclusion, three links from which you can learn more about Caffe and Deep Learning:



UPD In the comments indicated that the three avatars are not cats. So the method finds 745 cats.

Source: https://habr.com/ru/post/231699/


All Articles