By cropping a photo to its most salient part, we can use Google’s “search by image” feature to identify object in a photo that does not have accompanying texts, caption or metadata.
At times, a web researcher or online editor needs to identify unknown objects in a photo – logo, building, location, etc. Google has a “search by image” tool that compares the image with similar images on the web and suggests the best match.
To effectively reverse search an image, we need to crop the image to its most salient part; i.e., remove unrelated objects and background. Let’s look at two sample photos, attached below, that do not have any accompanying information; we need to find out where they were taken.
As explained in Google Help, there are four ways to search by image: drag and drop, upload an image, copy and paste an URL for the image, right-click an image on the web.
Let’s assume someone emailed me the two sample photos and I saved them on my computer; so I uploaded each photo to Google and tried searching by image. However, Google was not able to find a match for either one.
I then cropped out the adjacent buildings in the first photo and ran the image search again; this time, Google suggested the building is Chinese Theater in Los Angeles – and it is correct. Below are the cropped image and a screenshot of search results.
So, reverse image search may look easy; however, for the second unknown photo, cropping alone did not work. Below is a cropped version and Google was not able to find a match.
The problem is with the unrelated objects in the background, especially that highrise building. So, I did some quick Photoshop editing: erase the unwanted background and then crop again to remove the blank area. This time, Google correctly suggested this is Pier Cafe in San Diego. Below are the edited/cropped image and the search result screenshot .
A final note is that after erasing the background, you must crop to remove the extra white area surrounding the main subject. In my test of the Pier Cafe photo, I erased the background but did not crop out the blank area surrounding the cafe, and Google was not able to find a match.
Related posts:
One Response to Reverse image search: A Google tool every web editor and researcher should use