Wednesday, October 28, 2015

Search Challenge (10/28/15): Fake or real? How would you know?

Is it fake? 

... is a question that comes up fairly frequently when you're reading online content.  We see news stories, social media posts, or images that flash by and make us wonder "Can that possibly be real?"

This week's Challenge is to explore a couple of ways to check if something is real or not.  This came up for me this past week when a friend posted the following image on his/her Facebook page and wondered how CalTrans was going to get all of that mud and debris off of I-5. 

I'm originally from LA, and there HAVE been recent mud slides in the LA area on Interstate 5 (I-5) at a place called the Grapevine (so-called because it's a very twisty freeway through a narrow mountain pass).  I've driven that section of road a lot in my life, and this just doesn't look like the Grapevine.  

So I wondered--is this photo real?

This is what started me thinking:  What methods CAN we use to tell if something is true and correct?  What would you do?  

1.  Is that image above really from the recent (October 16, 2015) mudslides on the Grapevine (I-5) near Los Angeles?  

2.  How about this next photo?  Is this real?  

3.  And what about this one?  It looks so simple, but which of these two pictures of a piece of cake is the original? (The top cake pic, or the bottom cake pic?)  

As always, please explain your reasoning and methods.  HOW did you figure out if it was true... or a fake?  

Next week we'll continue this theme of "how to find fakes" with another kind of data!  Let me know how you like these Challenges.  I hope you find them as fun (and useful) as I do.  

Answers (and how I figured it out) next Monday, November 2, 2015.  

Search on! 


  1. A1 - Use Google image search. For the third result, it says it's a landslide in Keelung, Taiwan in 2010. Google 'Keelung landslide 2010'. Can quite confirm about it.
    A2 - Again, use Google image search. The first result links to . I think this press is okay reputable. So I think it's valid - . Normally, if I really want the answer, I would write to the editor of the article. But you may have many readers doing it, I am not going to bother him this time.
    A3 - I would say the bottom cake pic is original. The fork has white border when zoom in.

  2. Good Morning, Dr. Russell and everyone.


    First, read the names you gave to photos to search better.

    1. Mud slide. 2. blob fish 3. cake NMD 4. just cake

    [Identify fake pictures]

    How to find out if a photo your friend posted online is fake

    13 online tools that help to verify the authenticity of a photo

    First one is easy at first sight because I don't think trees will be that way, besides California has been on drought so to much green. To verify I tried Search by Google Image.

    [Interstate 5]

    Mudslide buries I-5 north of Los Angeles in 5 feet of mud

    2. It is real. Search by Image with Google says it is really blob fish.

    3. Look for data using some tools that articles mentioned. Also, think that NMD could be kind of signal. Besides, fork doesn't make shadow.


    1. Is that image above really from the recent (October 16, 2015) mudslides on the Grapevine (I-5) near Los Angeles?
    A: No. They are from Taiwan 2010

    2. How about this next photo? Is this real?
    A. Yes, it is The grumpy-looking, gelatinous blobfish has won a public vote to become the official mascot of the Ugly Animal Preservation Society.

    3. And what about this one? It looks so simple, but which of these two pictures of a piece of cake is the original? (The top cake pic, or the bottom cake pic?)

    A. With Fotoforensics, both images looks like the cake was added after. Bottom cake is the original.

    1. Nicely done Ramón - good finds on the tools - useful list from the comments in the API link you found:
      user name: RobinGood
      looks real fake to me

    2. Thanks Remmij. Sorry for the late message. I had some situations. Yes, a new Turtle. Courtesy of Dr. Russell. Looks amazing right? "My" Egret will not come this year and therefore, needed new avatar.

      I always enjoy your links and photos. In fotoforensics photo you created, I notice that color looks like was added later and grey or black no, so maybe I need to learn more about that because many parts of the photo looks real.

      How did you find those cakes? I'll look the Robin Good document, looks very interesting and super helpful.

      Also as Luis mentioned, I knew that edited photos have more MB than original ones. And, maybe that is wrong because if we save it with another file format we can change that, right?

      The Challenge made me remember this The Verification Handbook and now that visited again, it is available in more formats and more languages.

      About "Trust but Verify", I just recently learned it comes from Russia. When I hear that was a great surprise.

    3. That is right, Remmij, Cats and goats are always present and more because yesterday was International Cat Day!

      Is this real or fake?

      Start Using Google Image Search to Kill Fake, Viral Pictures

      I also forgot to ask you, how did you find Robin Good document?

    4. "I also forgot to ask you, how did you find Robin Good document?"
      Ramón - it was in the comment section of the API link you had in your October 28, 2015 at 9:46 AM post [one comment - RobinGood - has link]
      …real or fake? certainly more plausible than this ⬇,
      flor de tortugas marinas
      but I couldn't confirm its authenticity - did find the image used at other places on the net, but nothing that identified a credible source…
      IMHO, I would doubt a hummingbird would choose such a precarious, albeit lovely, place to repose or that the blossom would support even the minimal weight in that way…??

    5. Thanks Remmij! I didn't see that part. Your flor de tortugas is very pretty, I like it!

      About coordinates in the cake photo. I didn't saw that either when trying metadata. I wonder if coordinates stay once you edit a photo. I guess no, right?

      Happy Halloween

    6. this challenge has been interesting - plausibility detection continues to be an issue… went back & found a slide in Nova Scotia that seems similar - cape islanders/novis & Acadian vessel
      seem to point toward Cape Breton, but who knows? Maybe some tunes will clear the ol' noggin…
      Mark's enhanced message… cat, goat, Dan (@fb -what the heck?), groundhog, Crab Nebula, etc.
      Cape Breton landslide, Sheet Harbour, NS… Pallas & marmot bombing
      …is one more real?
      landslide, pumpkins
      landslide, nicks/buckingham

  3. A1: Used Search by Image. Dragged and dropped the picture in the search box. Chose "All sizes" First image is from The Guardian page of 25 April 2010, stating that it is in Keelung Taiwan. The Guardian is a newspaper with a decent reputation. SO Not in or near LA
    A2: Same approach. 5th link on the result page was from National Geographic. Fish is called "blobfish".
    A3: Same approach as previous ended in no results. So I zoomed in and saw a white "border" in between the fork and the napkin. It is hard to remove reflections, shadows etc from pictures. So my guess is that the bottom picture is fake.

  4. A1: Used "Search by Image". On the result page I chose "Show all sizes". The first picture was from a page on The Guardian, a well respected UK newspaper. Said it was a landslide In Keelung Taiwan. So for me this is not in or near LA
    A2: Same approach. On the search result page the fifth hit was a page from National Geographic. The fish is called "blobfish".
    A3: Same approach but didn't bring up anything useful. So I zoomed in and saw a white "border" around the fork where it touches the napkin. I know that it is very hard to remove reflections, shadows etc when you try to manipulate any photos. So for me the bottom picture is the real one.

  5. I made a typing mistake. Wanted to say that the bottom picture is the REAL one.

  6. I made a mistake while typing. The bottom picture is the REAL one

  7. This was almost too easy. For the first one, I did a right mouse click to search Google for this image and found that that this was a picture of a landslide in Taiwan in 2010. Then I went to save the picture of the fish and it was labeled Blobfish, so I just Googled that and found that is is a real fish that lives in the waters near Australia. Although I don't have confirmation I think the bottom picture is the real one because it looks like the fork and the crumbs have been erased. Only took about 5 minutes.

  8. trust, but verify… the sources
    pic real, location not…
    Taiwan, Highway 3
    blog 4/26/10

    Trufflus Psychrolutes marcidus
    not to be confused with Bobthebuilderfish

    a lemon variety
    Raspberry & Passionfruit Mousse cake
    new Munch Museum coming in 2018?
    caloric scream in Oslo from your recent Norway visit?

    is Munch's Der Schrei der Natur real?
    "One evening I was walking along a path, the city was on one side and the fjord below. I felt tired and ill. I stopped and looked out over the fjord—the sun was setting, and the clouds turning blood red. I sensed a scream passing through nature; it seemed to me that I heard the scream. I painted this picture, painted the clouds as actual blood. The color shrieked. This became The Scream."

    what is fake? a couple variants:
    oui, Virginia
    RDF, Jobs

  9. Landslide pix is real. Image search took me as expected to Snopes Fauxtography. It happened in Taiwan April 2010.

    More at


    The blobfish Psychrolutes marcidus is real. Wikipedia article cites many references. Found by Image search again.

    The ice scream pie: jury still out. Likely devouring it.

    jon tU who thinks these are great Challenges

  10. it all must be true… it is on the screen, right in front of me… and the pictures are moving - with music and sound - all very moving…
    b***fish vid
    You Tube as a source
    10 list
    shark tales
    (worth the watch to see this:)

  11. Image search for the first two pictures gives the result immediately.
    On the third one the levitating fork doesn't cast any shadow. Fake. (almost a photoshop disaster)

  12. I have spent a good 2 hours with Fotoforensics, which I had never heard of til today. Its my conclusion that they are both unaltered original pix. FIrst pix without the fork snapped, then a little chocky smudging and shiny fork dropped in and another original is snapped.


  13. 1. The Guardian's photo (on their already mentioned page from 4/25/2010) has the following subtitle: "Keelung, Taiwan: A landslide engulfs National Highway No 3 / Photograph: Reuters". However, I can't find the picture on Reuters Pictures. I still don't know why. Highlighting and right-clicking part of the subtitle, I can Google Search [ Keelung, Taiwan: A landslide ]. Wikipedia is one of the search results. Its article, however, directs to a broken link. Translate landslide to Chinese gives 滑坡. Searching images with this string, limited to dates between 4/23/2010 and 4/30/2010, gives a vast amount of Chinese sources confirming the subtitle. Weirdly, a similar video search produces nothing. I wonder why Google Video Search works so bad — all the time, in fact, to the point it's extremely rare for me to find anything on Google Video Search I couldn't find on YouTube. The exact place is easy to find on Google Earth, where two Panoramio photos confirm it too.

    2. At the time of the "ugliest animal contest", I had already conducted this research and found it to be true. It might be of value to add that the blobfish's blobbiness is increased when we see them, because they're much less blobby where they live in the deep waters of the ocean. One credible source, for example, is the SciShow Kids.

    3. To me, both pictures look shopped, in fact. The white "stains" where the crumbles are missing seem like when you use the clone tool on an image manipulation program or website like GIMP or pixlr (I've tried the clone tool or, for the matter, any of those programs/sites very few times). The fork, though, seems to be floating. And there's a weird "bycicle-wheel" of light rays near the fork head that I can't explain.
    I haven't seen this mentioned anywhere but I suspect that a photoshopped image is saved in a larger file than the original. Unless we take out a lot of features and the amount of similar pixels increases, this is probably true, but it's only a guess from someone who knows close to nothing about this. The second image (no fork, no crumbles) is smaller so, if my hypothesis holds true, it might be an earlier version.
    I also found the same Yellow-5 cheesecake as remmij's (I don't believe it's real lemon…) on a Japanese blog post by a lady who names herself chocolalait (or Shokorare, as Google transliterates her Japanese name ショコラレ). She says those photos were taken on the café at the entrance of Munch Museum in Oslo. The plate could be the same and the fork is quite similar (although probably not exactly the same design). Coincidentally, there's also a Munch Café in Tokyo serving a similar type of cake.

    1. Nice find (of the Japanese woman who also photographed the Scream chocolate on the cake in the museum)

  14. LMV - nice find on the Munch Café in Tokyo cake — I see yellow custard, I think lemon - Pavlovian.
    hopefully Dan does a small epicurean review.
    - did you see the sushi version?
    azo dyes

    1. I know what you mean, remmij. I see bright-colored glossy plastic-looking food, I think E numbers. It's this kind of Pavlovian, I guess. (I'm lucky to having been born in a time and place where food additives were frowned upon, and having had a mother who was a both a concerned doctor and an amazing cook. So my tastebuds were used from childhood to traditional recipes and I guess that's why I still dislike artificial flavors after all these years and food experiences.)
      The sushi version is cute.

      After the Munch Museum has turned a disturbing painting into a supposedly attractive cheesecake, I wonder if the Prado is planning to present some desserts inspired in Goya's Saturn Dvouring His Son. At least the subject is already food-related…

    2. Wow, the NY Times article is really interesting!! (and also nice to compare to the official page on the Prado's website). This reminds me of the amazing detective story around the invention of eyeglasses. I wrote a chapter on the subject for a forthcoming book and I'd recommend Chiara Frugoni's entertaining but accurate Books, Banks, Buttons, and Other Inventions from the Middle Ages as an introduction to the complexity of the whole matter (its first 6 pages are available on Amazon).

      The Saturn/Scream amalgam is a gem!

      And thanks for the links, reminding me why I've always loved Goya so much.

    3. thanks for the Prado & Chiara Frugoni points - will look at them
      …this may be a bridge too far, but I lack discipline… apologies to y Lucientes & Saturnus & forks everywhere -
      alternate Saturn menu
      "San Francesco e il lupo."

    4. The black paintings have always amazed me... And terrified me!

  15. #1 This is when I want the story to bring me to the photo. Yes there was a massive mud on the Grapevine on Oct 16 2015 which was reported by CNN, News Today etc. The images in the news are dramatic but not to the degree the image in question would suggest.

  16. The cake images require an examination of highlights, shadows, perspective and relative dimensions. There are two distinct differences. Crumbs and the fork. I think these differences are throwing us off track. There is something about the perspective that doesn't sit well with me. Relative sizing as well seems off. So my answer is neither.

    1. Sitting in one of my favorite cafes I decided to do an experiment. I took two pictures of some chocolate squares. One from a side view and one overhead view. Notice the perspective of the plate and the chocolates. An overhead shot can't capture the side of the square. But in Dr. Dan's images we have an overhead shot of the plate with and side view of the cake. The perspectives of the items are not on the same plane IMO.

    2. Rosemary - did you see these brownies too? with Dan's tasting forks… spooky, how they seemed to morph right in the case ;)
      (thanks to Eric for finding the requisite fork - good filter search)
      it may be a lens thing… focal length, depth of field, lens angle/perspective or other optical lingo/physics?

    3. Another aspect of this cake that popped up is the geolocation 59°55'02.8"N 10°46'30.1"E puts it near the Munchmuseet and the famous "The Scream" by Munch which we see as decoration in our images. Curious. So we have The Scream Munch Cake. Image search "the scream munch cake" and we get lots of hits. This is a twist I wasn't expecting. So course I am now wondering how Dr. Dan chose this picture.

    4. And here we have the same cake with spoon at the Munch Cafe in Oslo.

    5. Well spotted! After I found the yellow cake, I searched [ skrik munch kafe kake ] and similar strings, hoping to find an official depiction, but forgot to do the much simpler search [ scream munch café cake ] or similar, in order to find other photos taken by visitors… ��

    6. Rosemary, after you pointed out the top view of the plate and side view of the image, I looked more closely at both the images where the cake and plate meet, and also where the table and plate meet. I noticed that there is a very strange reflection of light. they sort of look like the whole photo was taken through a window, but I think it is a sign of photoshopping. The crosses over the plate and table with out curving to match contours of the plate and table.

      Also, in the bottom right corner of the image, if you look at the shadow cast by the plate on the table plate, the shadow of the plate has a firm white-ish outline, rather than the sort of blurry line that shadows normally cast.

      This seems like clear evidence that both images were photoshopped. I think this is image is made up of separate plate, table, cake + napkin, and in one photo fork images.

    7. Rosemary, could you please tell where/how you got the geo-coorindates (59°55'02.8"N 10°46'30.1"E) from the cake photo(s)? can't seem to locate it…
      as you know, Dan was recently in Lillehammer - probably transited thru Oslo and took the chance to visit the Munch Museum & Cafe -
      he has an eye for art and a sweet tooth so I believe he snapped & altered his cake photos from that recent visit.
      not all the garnishes are highly detailed
      in mass
      different presentation, same style fork, 4 tine variety
      nearly after
      group scream

    8. Remmij because of your question you may have pointed something out that I had not noticed. I use a Chrome extension EXIF viewer. So in the second cake image I have a little camera icon in the bottom right corner. I click on the camera & it opens showing the data, histogram and geolocation. The first image doesn't have these details which would suggest if there is a legitimate image it is the second image. The first image has limited data, histogram & no geolocation. I think you´re right Dr. Dan gave himself a nice treat after visiting the museum and snapped a picture. I still question the perspective of both however.

    9. Cake First Image Data "Date Time 2015:10:28 05:23:08"
      Cake Second Image Data "Date Time Original 2015:10:12 06:14:47"

      Dates are 16 days apart and Second Image states 'original' but the first (top) image just says date and not camera but 'Google'. If we tracked Dr. Dan we would probably find he wasn't in Olso on the 28th.

    10. Rosemary - thanks for the info on Chrome extensions, had tried Jeffrey's Exif Viewer and the inspector in Safari (my regular browser) and failed to get any geo info…
      will look into adding an exif viewer (Aruna or Rodrigue) and trying a look with Chrome… may have to mix up my browser usage - using it on both photos is a clever
      way to distinguish which was altered - good catch.
      Having access to the geolocation coordinates is a handy piece of the puzzle, again thanks! Curious that some exif viewers can extract them, while others can't - perhaps Dan will explain?

    11. Rosemary, added this exif viewer to Chrome & it worked as you said - terrific help!
      the images associated with the map even included cake and expresso - and a NEW fork form - and art to boot… thanks again for pointing out this tool.
      thought the 3 tine fork was interesting
      even paintings off maps - great!
      a slice of Munch wall
      (even IDs the camera… still don't know why Jeffery's doesn't seem to extract the info???)
      Motorola 2013 Moto X XT1058

    12. Blobfish search at NOAA National Oceanic and Atmosphere Administration provides these images and video of the Blob sculpin (Psychrolutes phrictus) which look much like the image we have here.

  17. This comment has been removed by the author.

  18. #3 -- I had some recollection that you could figure out if an image had been photoshopped by analyzing the metadata, so my first step was to look at an online exif viewer, but it showed near identical information for both photos. The only difference between the two images was in a field called "RenderingIntent". In this field, one image had "Media-Relative Colorimetric", and the other "Perceptual".

    I considered looking up the difference between these, but instead I googled [Detect if photoshopped]. The top result brought be to, and i uploaded both images. The results told me that both had "been opened and resaved with a program", and had "Adobe editing tags", but it also cautioned that "Exif and image data can tell if an image has been edited, but not how much."

    I looked back at the images, and I am of the belief that the original photo had both cake smears and no fork, but one photo edited out the cake smears and crumbs, and the other added a fork. However, if I had more time I would open up both photos in Photoshop and zoom around the areas I thought had been edited to look for clearer signs of editing. I would also try out some other online metadata tools, to see if there is any other information that can be yielded from these tools.

    I tried to quickly check my hypothesis that there was an original photo, by using reverse image search, in hopes that different versions would be online, but that didn't yield any results.

    1. fwiw - to see how worked, tried it & this was the result:
      iffy test
      but, searching "really" brought up this site… nefarious entrepreneurs testing the parameters…
      the commerce of fakes -

  19. Another method for the cake.
    I searched in Google Image for "fork", with search tools set to color=transparent. As third hit, I found this very same fork. With its transparent background it could easily be copied over the (apparently original) cake picture (which was not found in Google - others than on this page).

    1. I didn't know you could set the colour to transparent. neat!
      I tried looking for cakes and plates with the background set to transparent, to see if the whole thing is a composite of different images, but so far haven't found the plate or cake in separate images.

  20. Curious to me that only Ramon and self have used Fotoforensics on the Scream. I'd sure like to see what other people figure out using it, especially the ELA feature

    Cheers jon

    1. analysis seems to take some skill, experience & interpretation - not just a click process - common sense and real world experience are good starting points -
      the landslide question is a good example - there is no no manipulation of the photo, but a misattribution of the location associated with the photo.
      There is also knowing the original source of the photo - a photo could be willfully discredited through a small photo edited change - for example someone catches
      a record fish & photographs it as proof - a competitor photo edits something in the photo and widely distributes the picture on the web - so that when it is analysed the
      photo appears altered and then is discounted as a fake even though the basic claim remains legitimate - classic mis-information tactic to introduce doubt.
      detailed discussion
      pretty good starting observations from Quora
      frm FotoForensics
      an interesting alternative with examples
      ghacks from a few years ago
      FF tutorials

    2. I agree, Remmij. I liked trying Fotoforensics as Jon mentions but I don't think that looking I can decide if photo is real or fake. In some cases, yes, but mostly no, as you showed us with your photos. You proved that some fake in other software were marked as real and viceversa. Real or fake is specially hard when someone like you creates it.

    3. The Iranian missiles story is an excellent reminder of the importance of understanding how images can be hacked. (And if you can do still images, you can do video as well...)

  21. would the indication that this image has been "photoshopped" (exif data) diminish its credibility or make it less 'real'? Discretion still required.
    Gates image

    fwiw, the UK style chalk drawing at the landslide site… the waters further muddied.

  22. finally located an image of the I-5/Grapevine slide… sorta
    this has been fun - pushed me to think out of the box and the sRs folks made some interesting finds and new tool discoveries…