Picollator FAQ
Why Picollator?
Picollator is the only hybrid search engine that uses photos and combined query as search criteria instead of pure textual queries. Its concept is to deliver best search results using flexible requests.
How does Picollator work?
Do not care. It works.
Recogmission team made significant efforts to simulate mathematically how people represent the environment in their brains and how people connect text to the visual environment. We use complex technologies that enable Picollator to distinguish different objects in the pictures, recognize and index them along with textual data, holding patents on that.
How Picollator searches for people faces?
Picollator processes images found in the web, detects faces, recognizes them and classifies as if it does the alphabet or dictionary. So next time you submit a photo Picollator searches through the existing image index and finds the picture similar to your picture.
What does Most popular sign on the Home page stand for?
Picollator shows Most Popular searches for photos with people on the Home page. You always see 5 pictures that appear in the search results very often.
You can also see results for Most popular photos. Just click the Similar images button under the image.
How does Similar images option work?
Picollator links visual content which is similar to each other to one cluster. So next time you click on the Similar images button under the image, you will see other pictures that the system considers similar to the one.
Do people peculiarities (moustache, hat, eye glasses) influence image search?
There is no system that can define at 100% who is the man. Remember that even people mix up faces or cannot recognize an acquaintance.
So, if Picollator cannot recognize who is in the picture, the system tries to give some possible variants.
Why did my query return no results?
Some of the potential reasons are that your photo does not resemble to anyone in our existing index, or the uploaded photo is not good enoug. Read about photo requirements in this FAQ. Please take into account that Picollator is now at the beta stage. The index is growing, but it has not covered too much of the Internet. Try texts together with images or just the text.
I submitted the URL, but Picollator did not find pictures. Why?
In the existing Picollator version, URL is the path to the image itself. Even if you can see the image on the screen and copy its link, it does not mean automatically that it is the real image. For instance, if you copy links from the first screen of Google search, those links do not contain image urls. You have to click on any thumbnail, and then select the url of that image from the top of the page.
Why I upload a photo with a woman and obtained results with men’s photos as well?
People of different gender may look like each other. Children may look like their father or their mother. That is quite normal.
What kind of image I have to submit to get better results?
Optimal image characteristics for best results are:
- less hair across the face;
- no glasses;
- a close-up of a face;
- avoid head rotations;
- image size: 1200х900 pixels;
- face size: more than 240х240 pixels.
What limits for images to upload?
Practically, Picollator is able to process any typical image. However, to ensure speed and results, please, do not upload images that exceed the following characteristics:
- maximum image resolution: 25 MPixels;
- minimum image resolution: 140x140 pixels.
What is the difference between Web search and Image search?
Image search enables you to search for all similar photos of people in the Internet by text, sample images or hybrid queries containing both data types.
Web search enables you to find web sites with the data matching your query. So in Web search mode you obtain web sites ranking instead of photos in Image search.
What does the button 'More results' mean?
If you press this button, you receive resources from this site stored in Picollator index and matching to your request.
How are search results in Web search ranked?
Picollator web search results rating is based on unique hybrid algorithms by Recogmission. They consider text-to-text matching as well as text-to-image and image-to-image similarity ratios. All results are listed by decreased match to your query.
Which browsers do Picollator Universal Search support?
Picollator supports MS Internet Explorer 6, MS Internet Explorer 7, Opera 9, Firefox 2, Firefox 3, Safari 3. We recommend using MS Internet Explorer for the exiisting version of Picollator.
What is the little tomato in the right corner of the uploaded photo for?
You have uploaded an image as a search query, but you do not like the results or you want to search for something more using text query only.
To disable the image you have uploaded, press the tomato sign. The photo and the tomato will turn grey.
Picollator will not pay attention to this photo anymore and search only within text index. Happy and safe searching ;)
What is the little tomato and text under the photos in search results stand for?
Great, that you are trying to use all possibilities of Picollator search. You uploaded an image and also added a text search query. Right?
To define which photos were found by similarity and which by the keyword you entered Picollator uses these signs: tomato stands for similarity and “text” sign stands for text.
Of course, you can have photos signed by both icons, these are found by both, image similarity and text.
How to work with Picollator plugin for Internet Explorer?
Picollator plugin adds an option “Picollate it” to your Internet Explorer contextual menu.
You see an image on a web site, you want to find other pictures similar to this in the Internet. Just right click on the image, then choose "Picollate it".
Picollator plugin brings you back similar images search results from Picollator index.
Install Picollator plugin.
What do 'Collapse all' and 'Expand all' links perform?
'Collapse all' mode forces Picollator to group images by the content meaning. As a result, you can obtain same list but consisting of groups corresponding to different people. For instance, if you submit a text query like 'anna', first group may contain only Anna Paquin, while another set is about Anna Faris. This feature is definitely unique, because other engines cannot understand which pictures should be set together. 'Expand all' turns back to the plain list of results.