Pre-Word

Download Version 3 of the GPRSG. It is now available and now comes with GPRSG Duplication Image Finder Software. It gives home users the ability to search for objects inside pictures, true picture recognition technology.

There is now a trial version of the Duplication Image Finder Software. The Duplication Image Finder Software is free when you download the General Picture Recognition Software. The duplication finder software is a good little utility. However it is not picture recognition software that works completely differently and is very powerful indeed. The general picture recognition software is UNIQUE there is nothing available on the Internet that does the same job in the same way. The general picture recognition software will try to recognise the main object. The general picture recognition software attempts to achieve true picture recognition.

The question is can computer find any object within any picture and identify that object, at the moment the answer is NO it is impossible to achieve this because the human brain is just too complex.

If computer cannot achieve 100% accuracy in object recognition how accurate can they be. This really depends on a lot of things it can depend on the object itself it can also depend on other objects in the picture it also depends on distance, background etc. The parameters are large. Therefore some objects and some pictures are easier to identify than others.

To find a duplicate picture is easy (please try my duplicate image finder software). Duplication image finder software is relatively easy to design and build. They look impressive and actually do a good job of identifying duplicate images or other files.

Duplication image software is good at identifying duplicate files but no good at identifying similar pictures or images. If there is a very small difference between pictures then it looks very impressive but actually can only be achieved if both pictures have very similar attributes. The problem with duplication image finder software is that after this very narrow margin pictures can be very different. This makes image duplication software useless for finding pictures with say the same car but with two very different backgrounds.

To achieve this we need something a lot more complex. What is required is true picture recognition software. This is the goal of my "General Picture Recognition Software" it tries to attempt the almost impossible task of finding similar objects inside pictures. For a full explination of its complexity please see below.

Professional Version is for business only and can only be downloaded after proof of business is made available.

Research Version Now Available for Download

THIS SOFTWARE WILL ACT LIKE MANY ROBOTS BUT INSTEAD OF SHARING ITS ENVIRONMENT INFORMATION IT WILL SHARE WHAT IT LEARNS ABOUT PICTURES IT HAS ANALYSED WITH ALL OTHER USERS.

THE MORE USERS THE GREATER ITS ABILITY WILL BE TO LEARN AND SHARE ITS KNOWLEDGE WITH ALL OTHERS.

PICTURE RECOGNITION AS NOW TURNED A NEW PAGE.

• True Picture Recognition Software that attempts to find the main object within your picture
• Sophisticated Image Search and find software with built in organiser.
• Will try and find the same object but in pictures with different types of background
• Finds similar picture scenes, say of a beach, lake, mountain range etc. using landscape mode
• Finds identical duplicated images fast (but unlike duplicate image software it will attempt to find objects also)
• This software also now comes with GPRSG Dedicated Duplication Image Finder
• Built in easy to use New Picture Organiser that allows you to organise pictures quickly into folders
• But most importantly this software Finds those pictures you never knew you had on your computer
• Networks, External Drives, Hard Drive, External Devices like, Cameras it can search most devices.
• It is impossible to look through 50,000 pictures but let this software find the pictures you require

Find different types of flowcharts etc. see picture below. This software attempts to find pictures of the same type of group, scene or object whether its flowcharts, forms, cars, shops, scenes of a mountain range, beach or lakes. This software can be used to find view and even organise those pictures on the computer using PICTURE RECOGNITION TECHNOLOGY.

• The only software with REAL and AUTONOMOUS thinking (it really learns from previous searches)
• Uses analysis information gathered from other user searches AROUND THE WORLD using QUARK FILES
• Internet Quark files can be turned ON or OFF if off only local analysis quark files are used
• Use what it has learnt from other searches made by other uses AROUND THE WORLD to improve your search
• Real thinking allows it to learn from user activity
• Autonomous Thinking allows it to learn itself.
• The software ATTEMPTS TO NAME THE MAIN OBJECT IN YOUR PICTURE using REAL THINKING
• All this provides for higher accuracy in its main task of recognising objects within pictures.

Version 3 now has real thinking and autonomous learning built into the AI part of the software. This makes this software very powerful. Using Quark Internet Files THE SOFTWARE CAN USE WHAT IT AS LEARNT FROM OTHERS WHO HAVE MADE PREVIOUS SEARCHES FOR SIMILAR OBJECTS AND HELP YOU FIND THOSE PICTURES YOU ARE LOOKING FOR ON YOUR COMPUTER, NETWORK OR BACKUP DRIVE.

In wide mode it will pick up lots of similar pictures not just very close matches. Therefore it will pick up pictures, that are not close matches in say position but still have lots of the same things in them that maybe worth having a look at. With the general picture recognition software the search can be narrowed or widened depending on requirements.

WHAT THE SOFTWARE DOES IN NON TECHNICAL TERMS

THIS SOFTWARE CAN EASILY FIND IDENTICAL IMAGES, BUT IT CAN DO MUCH MORE, THIS SOFTWARE TRIES TO FIND OBJECTS WITHIN PICTURES AND A WIDER RANGE OF PICTURES THAT ARE SIMILAR TO THE ORIGINAL PICTURE. THIS IS TRUE GENERAL PICTURE RECOGNITION SOFTWARE.

IF YOU ARE LOOKING FOR SAY ALL THE CARS ON YOUR HARD DRIVE DUPLICATE SOFTWARE CAN ONLY FIND A PICTURE OF A CAR THAT HAS A VERY SIMILAR BACKGROUND. PICTURE RECOGNITION SOFTWARE TRIES TO FIND CARS THAT ARE DIFFERENT IN SHAPE AND COLOUR AND BACKGROUND. THE SOFTWARE CAN LEARN FROM OTHER USERS AROUND THE WORLD AND THEREFORE THE MORE IT LEARNS THE MORE HITS IT WILL GET AGAINST OTHER TYPES OF THE SAME OBJECT.

IF YOU HAVE EVER SAID I KNOW THERE IS A PICTURE OF THAT SOMEWHERE ON MY HARDDRIVE BUT CANNOT FIND IT, ANYMORE THEN THIS SOFTWARE IS FOR YOU. IT WILL FIND PICTURES YOU NEVER KNEW YOU HAD AND WILL SHOW YOU THE MOST RELEVANT PICTURES TO YOUR SEARCH.

APART FROM HARDDRIVES, NETWORKS, BACKUP DRIVES, ANY DEVICE THAT CAN PLUG INTO YOUR COMPUTER LIKE A MOBILE PHONE CAN BE ANALYSED BY THE SOFTWARE,.ALLOW THE PICTURE RECOGNITION SOFTWARE TO FIND THE PHOTOS YOU ARE LOOKING FOR SO YOU CAN SELECT THE ONE'S YOU WISH TO PUT ON YOUR COMPUTER.

THE BIGGEST INNOVATION IN VERSION THREE IS THAT THE USER CAN SELECT TO SHARE ANALYSIS INFORMATION WITH OTHER USERS AROUND THE WORLD (SEE BELOW). MAKING VERSION THREE VERY ACCURATE INDEED. IT EVEN ATTEMPTS TO TELL YOU WHAT THE MAIN OBJECT IN YOUR PICTURE IS USING THE REAL THINKING MODE

General Picture Recognition Made Simple and Easy to Use

To use the software you normally need only do three things!

1) Type the name of the object you wish to find e.g. Car, Bus, Train, House or Landscape Seaside, Lake etc, most of the common objects are already contained in the sample database however you can easily add your own.
2) Select the harddrive or folder or network drive or any other external drive you wish to search
3) Run the Software so it searches for the required Pictures.

That is it the software is very simple to use. You can either continue with other work and review the pictures it as found later or you can sit back and watch the pictures in full page mode. You can sit back at a distance and watch it go through all the pictures on your computer. You can even go out for the day and set the computer to go into STANDBY MODE once it completes its task.

You can run as many instances of the software as you wish, therefore you can search for several different pictures at the same time. Also in each instance you can use more than a single template. So the amount of searching that can be done simultaneously is only limited by the power of your computer.

If you wish to understand all the modes and technology and functionality available and options you can select then please read on. Although the software is simple to use the complexity behind the software is huge.

Please go to the download page from here

There is a very small licence fee to use the software it covers web space cost etc.

ALL ABOUT THE SOFTWARE AND TECHNICAL INFORMATION BELOW -

FOR THOSE WHO ARE TECHNICAL and would like to know more about how it works please read on if you just wish to use the software then download from the link above.

New Technologies and a New Age for Picture Recognition (Version 3)

THIS SOFTWARE IS AT THE CUTTING EDGE OF GENERAL PICTURE RECOGNITION TECHNOLOGY!

VERSION THREE NOW HAS A NEW POWERFUL BUT SIMPLE TO USE BUILT IN PICTURE ORGANISER FOR PICTURE FILES OR PICTURE DOCUMENTS FOUND, THIS ORGANISER IS DIFFERENT FROM THE NORMAL TYPE PICTURE ORGANISERS BECAUSE IT USES THE POWER OF PICTURE RECOGNITION TO FIND THE MAIN PICTURE THEME OR OBJECT:, YOU CAN NOW EASILY GROUP PHOTOS SCATTERED ALL OVER YOUR HARD DRIVE, BACKUP DRIVE OR NETWORK BY LETTING THE SOFTWARE DO IT FOR YOU- SEE BELOW.

VERSION THREE NOW HAS REAL THINKING TECHNOLOGY THIS MEANS THAT IT TRIES TO NAME THE OBJECT IN THE PICTURE AND DISPLAYS THE NAME OF THE OBJECT. ANOTHER INTERESTING NEW TECHNOLOGY INCORPORATED INTO VERSION 3 IS AUTONOMOUS THINKING . BUT AUTONOMOUS AND REAL THINKING ARE POINTLESS WITHOUT THE NEW QUARK LOCAL AND NOW INTERNET MEMORY that allows the software to learn via quark distributed analyse files via the Internet OR JUST USE YOUR OWN LOCAL QUARK FILES. The Internet function can be turned ON or OFF by the user but if on ALLOWS EACH USER OF THE SOFTWARE TO SHARE PREVIOUS ANALYSED INFORMATION WITH ALL OTHER USERS AROUND THE WORLD THIS SHOULD IMPROVE ACCURACY AND ALLOW A WIDER SEARCH OF THE SAME OBJECT OR THEME (NO PICTURES ARE EVER DISTRIBUTED OVER THE INTERNET, AND NO PICTURES CAN BE CREATED USING THE ANALYSIS QUARK FILES). SHARING AND LEARNING MAKES THIS SOFTWARE UNIQUE.

There are now three versions standard, research and professional. The main difference between the versions is that the research and professional versions allows the end user to exchange locations of similar pictures and the E-mail address of the person who knows the location of the same picture files within that organization. However the professional version is only available to organizations that can show the status of their organizations. For further information on the professional version see the link to the left on this page. However the standard version does everything the professional version does but does not allow exchange of picture file names and contact E-mail addresses. The research and professional version is ring fenced with further E-mail request protection. Therefore security is alway maintained.

THIS IS THE NEXT GENERA.TION OF PICTURE RECOGNITION

IF YOU ARE A UNIVERSITY AND ARE INTERESTED IN THIS AREA AND WOULD LIKE ME TO GIVE A LEC TURE ON THE SUBJECT AREA UK ONLY PLEASE CONTACT ME VIA E-MAIL. ALSO SEE BELOW UNDER RESEARCH AND UNIVERSITIES.

THE QUESTION IS CAN SOFTWARE NOW THINK WELL ENOUGH TO MAKE SENSE OF PICTURES?

GENERAL PICTURE RECOGNITION SOFTWARE WHAT IS IT

Unlike things like Face recognition that is specialized but can only be used to identify one thing i.e. a face, general picture recognition attempts to recognize any type of object from faces to people to cars to landscapes, castles, ships, coins etc. it will attempt to recognizes the main object within a picture whatever that object maybe. However unlike face recognition it could not tell if the faces were the same two faces in any type of picture. It may tell the user that two pictures both have a face in them but is not dedicated enough to tell if those two faces are of the same person, if you wish to achieve this you will need dedicated face recognition software. The main advantage with this general picture recognition software is that it can try and identify many different types of objects where dedicated face recognition only attempts to find a face.

Recognizing the same picture is not that difficult, the difficulty arises when there are two very different pictures but they both have the same main object or theme. For instance finding ALL pictures that have CASTLES in them seems easy but what if the CASTLES are different in colour or shape. Then most would say it would be impossible to find VERY DIFFERENT PICTURES but with the same theme. THIS IS ACTUALLY TRUE IT IS INCREDIBLY DIFFICULT TO FIND TWO PICTURES THAT HAVE THE SAME MAIN OBJECT BUT ARE VERY DIFFERENT. This is the main goal of the GENERAL PICTURE RECOGNITION SOFTWARE it attempts to do the almost impossible task of identifying very different pictures that have the same theme or object in them.

PICTURE AND INFORMATION OVERLOAD

The main point of this SOFTWARE is to REDUCE information OVERLOAD. This software searches for a certain type of picture and finds the best match for that picture and similar pictures on your computer. IT LOOKS THROUGH ALL YOUR PICTURES FOR YOU THEN reports back and provides you with the best pictures and pictures that are of the type you are looking for and require.

INTERNET AND QUARK MEMORY - NEW TECHNOLOGY

A quark is an undefinable particle science knows or believes it exists but that is as far as it goes. The general picture recognition software when it analyses a picture stores a very small amount of analysis information about each picture. For instance it will try and name the main object in the picture and store this information (PLEASE NOTE THESE FILES CANNOT REPLICATE ANY PART OF A PICTURE) they will hold things like Car as the main type of object in a previous picture analysed. THE MOST EXCITING PART OF ALL OF THESE QUARK MEMORY FILES IS THAT THEY ARE DISTRIBUTED BY MY WEB-SITE OVER THE INTERNET. So if you are looking for a Tree then the software can download a Quark file that has lots of analysis information from many other users about TREES. This information acts like HUMAN MEMORY it allows the software to use previously learnt knowledge and apply that knowledge to finding pictures with the same object on your hard disk, network drive etc.

THESE QUARK FILES WILL BE CREATED USING AUTONOMOUS AND REAL LEARNING TECHNOLOGY ALREADY BUILT INTO THE SOFTWARE. IN OTHER WORDS THE SOFTWARE WILL DECIDE WHAT IS A CAR, HOUSE, PICTURE OF THE SEASIDE ETC USING ITS OWN INTERNAL BASE PROGRAMMING BUT MORE IMPORTANTLY ALSO USING INFORMATION FROM PAST LEARNING.

USING THESE QUARK ANALYSIS FILES WILL BE COMPLETELY UP TO THE END USER THEY CAN TURN THE INTERNET FUNCTION OFF AND ONLY USE LOCAL QUARK FILES IF THE USER WISHES. HOWEVER IF THE USER DECIDES TO BENEFIT FROM THE USE OF THESE QUARK FILES VIA MY INTERNET SITE YOU MUST ALLOW YOUR LOCAL QUARK FILES TO BE USED AND DISTRIBUTED TO OTHERS VIA MY INTERNET WEB SITE. THESE QUARK FILES ARE GENERATED BY THE SOFTWARE AND ARE COPYRIGHT PROTECTED. THE INFORMATION IN THESE QUARK FILES ARE ONLY USEFUL TO THE PICTURE RECOGNITION SOFTWARE.

VERY IMPORTANT - THESE QUARK FILES CANNOT BE USED TO REPLICATE ANY PART OF A PICTURE IN OTHER WORDS THEY DO NOT HOLD ANY INFORMATION THAT CAN FORM A PICTURE. IT IS IMPOSSIBLE TO REPRODUCE A PICTURE USING THESE QUARK AI OBJECT FILES. HOWEVER BY USING INTERNET QUARK FILES THE SOFTWARE WILL HAVE A WIDE RANGE OF PREVIOUSLY ANALYSED DATA OF OTHERS WHO HAVE USED THE GENERAL PICTURE RECOGNITION SOFTWARE TO FIND OBJECTS WITHIN PICTURES. THESE FILES WILL ALLOW A MUCH WIDER RANGE OF SIMILAR CLASS OBJECTS TO BE FOUND. SO SAY YOUR INITIAL SAMPLE PICTURE IS OF A STEAM TRAIN USING THESE QUARK FILES YOU WILL FIND THAT PICTURES OF SAY ELECTRIC TRAINS WILL ALSO BE PICKED UP.

THIS WILL BE THE BIGGEST INNOVATION IN PICTURE RECOGNITION IT WILL ALLOW MILLIONS OF USERS ALL TO SHARE PAST ANALYSIS INFORMATION OF PICTURES. THIS IN THEORY COULD EVENTUALLY MAKE PICTURE RECOGNITION OF ANY TYPE OF OBJECT OR PICTURE 99.99999999999 % ACCURATE. IN OTHER WORDS LIKE HUMANS THE MORE LEARNING THE BETTER THE RESULT.

SIMPLER INTERFACE

There is a large amount of SAMPLE pictures so all the basic types of search have predefined samples. So if you are searching for say a HOUSE, SHOP, CAR, TRAIN, PEOPLE, BUS etc most GENERAL OBJECTS that you see everyday are already built into the sample database therefore all you need to do is type the WORD for example CAR and the software will then search for all cars on your hard drive. However for users who wish to create their own samples they can.

THE MAIN ADVANTAGE IN HAVING THE GENERAL PICTURE RECOGNITION SOFTWARE

1. Save hours and hours of time searching for pictures on your harddrive yourself
2. Let this software find those pictures for you by
3. Compare your pictures with built in Sample Pictures and quark information
4. Create your own samples to compare against other pictures on your computer
5. Use exclude so you only have to analyze a picture in a category once
6. Save pictures of interest that have been found
7. Search your other network computers and backup drives for pictures
8. View pictures as the software finds them sit back and see what pictures are on your computer
9. Will find hidden pictures on your computer
10. Run in task bar while you do other things
11. New in Version 3.0 now you can Organise Quickly Pictures found into a single Folder. For instance if you have photos of CASTLES located in lots of different folders and you wish to GROUP them altogether into a single folder you can now do this. The Picture Recognition will try and find pictures of say CASTLES then you can use the new organiser to quickly put all the castles found into a folder called Castles or any other name you wish to give the folder. This can be done also if searching external drives or a network for pictures. These are some of the different types of castles that have been grouped into a folder called CASTLES
    
12. 
13. THIS SECTION MAY ALSO INTEREST THOSE WHO ARE RESEARCHING PICTURE RECOGNITION
14. New in Version 3.0 -(I wish not to go into too much detail but this outlines the technology) REAL THINKING TECHNOLOGY the software will REALLY LEARN from USER DECISIONS and incorporates this into its AI OR (REAL THINKING MODULE) this allows the software to learn just like a REAL HUMAN would learn new objects for the first time. This incorporates AI and Picture Recognition together with RT. or REAL THINKING. The more it learns the better the results will be with pictures that have the SAME OBJECTS but are VERY DIFFERENT from the ORIGINAL SAMPLE. For example two castles that look very different but are still castles. This will happen more often as the REAL THINKING with AI learns. REAL THINKING will DISPLAY THE CLASS NAME OF THE OBJECT IF IT THINKS IT IS A CERTAIN OBJECT.
15. IMAGES TAKEN FROM YOUR HARD DRIVE, NETWORK DRIVE ETC IN Version 3 are analysed using something I have called AUTONOMOUS THINKING this is WHERE THE SOFTWARE MAKES ITS OWN DECISION BASED ON THE PICTURE AND STORES ITS DECISION LIKE A HUMAN STORES WHAT IT LEARNS SO THAT IT CAN BE USED TO LOOK AT FUTURE PICTURES ANALYSED. This means that the new pictures that are analysed are NOT ANALYSED USING PARAMETERS SET BY THE CODE ALONE BUT LEARNT INFORMATION FROM THE SOFTWARE ITSELF. WHAT MAKES THIS EVEN MORE INTERESTING IS THAT SOME OF THE DECISIONS THAT THE COMPUTER MAKES BECOME UNKNOWN TO ME WHO ORIGINALLY PROGRAMMED THE SOFTWARE. THEREFORE I WILL NOT KNOW WHY THE SOFTWARE MADE THAT DECISION UNLESS I COULD ANALYSE ALL THE PREVIOUS DATA THE SOFTWARE HAS LEARNT FROM. THIS WILL BE THE FIRST PICTURE RECOGNITION SOFTWARE THAT HAS REAL THINKING, AUTONOMOUS THINKING,AND INTERNET QUARK FILE TECHNOLOGY IN RELATION TO PICTURE RECOGNITION. The difference between REAL THINKING and AUTONOMOUS THINKING is that real thinking uses what it learns from user decisions while autonomous thinking takes NO HUMAN INPUT AT ALL making its own decisions. Error calculations allow for mistakes, ("just like humans make them"), from the autonomous thinking modules, therefore they do not impact on the accuracy. Therefore its assumed that autonomous thinking will not get it always right every time. Autonomous thinking is also protected by the base code and if the base code disagrees it will reduce the possibility of error.
    In theory REAL THINKING and AUTONOMOUS THINKING will allow you to find a class of pictures that have very little resemblance to the original sample but will still find pictures that are of the same CLASS for instance it will be able to find a CLASS of castles even if some of the castles have no real resemblance to the original castle SAMPLE. This will happen more and more as the SOFTWARE LEARNS USING REAL AND AUTONOMOUS THINKING. Also the more it learns the more often it will be able to NAME and LABEL the CLASS of the main object in the picture. THE MORE USERS THE GREATER ITS ABILITY WILL BE TO LEARN AND SHARE IT KNOWLEDGE WITH ALL OTHER USERS.
16. You can now use the New Small Mode in Version 3 this works with a single object on a clear white background if you wish for best results. But it attempts to compare the sample and a single object however small or big against the sample. The object or sample can have extreme differences in size but the software will still try and compare both objects. This is good if you keep a collection of objects that you have taken photos of each object but with big differences in size. You can therefore have an object that would normally be to small to compare with a very large but similar object and it will still attempt to compare and find the object. The Small Mode is designed for those who wish to find a collection of pictures that have been taken of a single object on a white background but where the object may have a large size difference because the photo was taken at different ranges. There are some other limitations with the small mode for instance the sample and compared object must have the same position or orientation. This works differently from all the other modes, where the other modes may compare differences in size but will not have the amount of certainty that this mode tries to produce.
17. Large Close Up .......................Medium Range Photo ..............Small Photo of same object
18. 
19. Picture recognition has great difficulty comparing vary large objects with very small objects because of the loss of the required pixels when attempting to resize a very small object and making it larger. It's actually easier to reduce or make an object smaller however the compare is less accurate as the object is reduced to a very small size. However if the small object becomes too small and loses it outline then it will be impossible to compare both objects.
20. In version 3 it shows you a box that is puts around the part of the image it is trying to analyse the most.
21. New in Version 3 Large Screen Picture Viewer - you can now watch each picture as it is analysed in a large picture view. So if the laptop or computer is at the other end of a room you can still watch it analysing each selected picture.

WARNING NOTICE

• Home networks and other networks - viewing pictures on another network computer please get permission. This software is very powerful and should only be used to search network computer with permission.
• This is now version three, may I thank everybody for their feedback and interest with version two.

This software uses picture recognition techniques to find objects inside picture.

VERSION THREE INITIAL MATCH MODES

The initial match modes are used to quickly find different types of images or forms. Certain modes will make it easier to find certain types of pictures or forms in picture format. The default modes are very good but there are different modes for different type of pictures that you may wish to find. For instance flowcharts can be found using the Forms and Monochrome initial match mode. You will usually find this is much faster than other modes for finding flowcharts, forms or dialog boxes etc.

• Slow Auto Off - this mode allows you to set the auto function off. You can either continue in this mode or choose any other mode that is not an AUTO function mode and it will not go into auto. However if you choose an Auto function mode then it will return to auto function. This is also true if the end user choose to Force Auto to be on.
• Slow - Will Analyse All Pictures (this mode is retained from version 1 and version 2.)
• (DEFAULT) Colour Quick Match 1 Auto - Finds pictures that are similar in colour (Very Fast) will go into slow mode to analyse a folder in more depth then return to (Colour Quick Mode 1 Auto) to search for other colour matches. In (Version 3) the default mode is more accurate and gives better results on the initial find.
• Colour Quick Match - (New in Version 3) this is the same as Colour Quick Match 1 Auto above, however unlike the Auto mode it will NOT go into auto mode once it has found a match or near match. In other words it will not check the FOLDER where the picture was found. Using the AUTO mode above it will check the folder in-depth, the downside with the AUTO mode is that it takes longer to complete the scan but the advantage is that a more detailed search will have taken place. The Colour Quick Match will therefore be faster than the Default (AUTO) mode.
• Full Match Only - This is new in version 3 it is near exact match of the sample picture its the quickest of all modes. Therefore you select this mode if you wish to find an exact match of the original picture only or a very very close match of the whole picture..
• Object Center Colour Match - this mode is looking for the colour of the main object that is usually situated in the center of the picture. Its actually looking at the center colours of the main object so if you know the picture you require has similar colours in the center of its object then this mode is very good. However it is not an Auto Mode therefore it can miss out not so near matches on the initial search however overall should be quicker than an auto mode if the Auto mode goes into Slow mode because there are large amounts of similar photos.
• Forms and Monochrome - This is the FORM and BLACK and WHITE MATCH, say you have a form and you filled it in and scan it into the computer this will find that form using a blank scanned in form. It will also find pictures that are not exact but are still very close looking background included to the original form, even if that background is not black and white say it is blue, (most of the background should be the same colour. If so this mode can be very fast. Black and White pictues are more easily found using this mode because like with most forms it is searching out the background colour. You can use this match mode if you have a very similar picture but it is not an exact perfect match.  If the picture is just a copy of the original picture use FULL MATCH MODE instead. This mode as just about no redundancy at all. It is also a very fast mode like full match mode, this mode can also be used for duplicate image search.
• Pattern Only - Ignores the colour and will only analyse the Pattern
• AI - this is based on AI only for first recognition.
• All the Quick Match Modes in Version 3 change to the Auto Quick Match Mode if a very close or Match takes Place.
• Force Auto - this is new in version 3 and forces all the modes into Auto mode apart from slow auto off mode. This mode when selected will put you into AUTO mode after a close match picture as been found.

Use Picture Type Modes

The software has been developed to act differently depending on whether your in Wide, Normal, Landscape or Close Object mode.

• Landscape mode - this mode is designed to find the same type of landscape as the sample picture for instance it ignores pattern recognition and objects and shape recognition most of the time.
• Close objects mode -(Narrow) this mode is best used if you have little background and your sample object has definable strong lines like the picture object you are searching. Shape and size should be near each other. Orientation needs to be the same for best results. But if the object or objects in the picture are very similar and there is little background then it will find them with very little redundancy. This is a NARROW match therefore the objects have to be quite similar or it will not pick them up. Good if the item is close up and very similar to the original sample without a lot of background or very plain background. In auto search mode it is very fast. In Form and Close Match Mode it is very fast.
• Normal mode - normal mode is when both Landscape and Close Object modes are turned off this is the default mode. It gives a wider search than Close Object mode and landscape mode therefore there is more redundancy but it does find other interesting pictures along with the type of picture you are looking for. Orientation DOES NOT need to be the same you can have one upright and another picture on its side and it will still try to find the object. Unlike close object and small where both pictures need to be in the upright position for best results.
• Small Mode - (New in Version 3) if there is a single object in the sample and it is on a clear white background and there is a single object that is being searched for in a picture on a white background and they are the same object but sized very differently then this mode will attempt to find large differences in sizes. Orientation must be the same. This mode is good for finding other items in a collection that is at different sizes but the same item or very similar. AI is switched off and only a single sample is allowed to be selected. This is good for a collection of items like different type of stamps if you collect stamps etc. However if you have different colour backgrounds for different items then you should not use this mode. Best results with clear white background and no shadow. So if you are taking photos then suggest using the flash on the camera.
• Wide Mode -(New in Version 3) this is the same as normal mode but will catch more pictures. This mode may increase redundancy. It will catch lots more pictures than normal, it is slower but maybe the only way of finding a picture that is quite different from the original sample. Always try normal mode before trying wide mode. Things like finding FACES work well in this mode, because faces can have large differences in colour, but again it will find lots of other types of pictures -this I call redundancy. Wide mode is very good if you do not mind it searching through and analyseing lots of pictures, will give a wide selection of pictures and can be quite slow depending on the match mode selected. Suggest using the wide mode with the default match mode for best use.

Very Customizable Software

This software is very customizable with all the modes and with the editable AI (Artificial Intelligent) side of things where you can select certain objects or important colours within your sample picture that you wish to find in similar pictures. You can even select several similar objects across several similar samples. This allows you to home in on certain objects of importance. Once selected it will learn from your selection and will remember it when using the same sample in the future.

There is a small £10.00 download license fee for one years use of the software for private use only. This helps me to continue to pay for web-space etc. If you are a business and wish to use it for business use please contact me to discuss the type of use and a fair price for the licence agreement for one years.

I have spent many, many, many, hours of continues work on the software, it is a very interesting hobby for me therefore as long as the £10.00 covers my web cost etc. I will continue to improve the software.

This software is Copyright Protected and all intellectual rights also belong to myself.

ALL WHO THOUGHT VERSION TWO WAS VERY GOOD THEN I HOPE YOU WILL BE EVEN MORE IMPRESSED WITH VERSION THREE.

AGAIN MANY THANKS AND PLEASE KEEP GIVING THE FEEDBACK.

IF THE SOFTWARE CAN LEARN AND USE ITS PREVIOUS LEARNING THEN IT IS STARTING TO SIMULATE THE HUMAN CONDITION.

ALL YOU NEED TO DO IS ENTER THE NAME OF THE OBJECT OR THEME THAT YOU WOULD LIKE THE GENERAL PICTURE RECOGNITION SOFTWARE TO FIND ON YOUR HARD DRIVE, NETWORK DRIVE OR REMOVABLE DRIVE AND LET IT SEARCH OUT THOSE PICTURE FOR YOU.

CARS, HOUSE, MONEY, TABLE, TREES, CARAVAN, TRAIN, STATION, BEACH, AND LET THE SOFTWARE DO ALL THE HARD WORK. THEN USE THE QUICK ORGANISER TO QUICKLY ORDER YOUR PHOTOS FOR YOU.

TEMPLATE PICTURES CAN INCLUDE ANY TYPE OF PICTURE THAT IS NOT ALREADY IN THE SAMPLE DATABASE OF COMMON SAMPLE PICTURES.

IF IT IS ALREADY IN THE COMMON SAMPLE DATABASE ALL YOU NEED DO IS TYPE IN THE NAME OF THE OBJECT YOU ARE LOOKING FOR ON YOUR COMPUTER, NETWORK DRIVE OR REMOVABLE DISK DRIVE.

• Security - Companies can check that the pictures on their computer are the type of pictures they would expect on their computers.
• Parents - again parents could use this software to check what pictures are on their computers, check that your children are not downloading unsuitable images. Today with the Internet you never know who is downloading pictures to your computer. For instance when browsing the Internet all pictures are stored in a temporary location on you hard drive. This software is ideal for checking what is on your hard drive.
• Photography, if you are a photographer with hundreds of thousands of pictures on your computer and you wish to find similar pictures that you have taken in the past, but cannot recall where you put them, this is the ideal software for that purpose, if you have pictues stored on backup drives then again this is ideal.
• Industry - any application that needs to identify very similar images or duplicate pictures. There are a huge amount of applications that save pictures to your hard drive. This software can check given an image if there are similar images on a computer hard drive. If you are trying to colour match certain aspects of the picture the AI module really will help.
• Scanned Documents in a picture format- it can distinguish between documents so if you have an exact document in an image (picture) format and you scan an identical document into you computer and create a sample it will be able to find that document on your computer or other network computers. It can even find the same image even if there are say added words to the same form. Therefore just because the form has handwriting on it, that should not stop it from being found.
• Organise and group your Pictures from ANY LOCATION ON THE HARD DRIVE and group similar pictures say of vehicles into one folder location. This has been added to Version 3 because of USER FEEDBACK.
• Big backup drive picture recognition search - some companies and many people make very large backups of hundreds of thousands of pictures. Impossible to search through your backup drive or drives to find pictures taken a long time ago this is the ideal software for searching out lots of pictures on backup drives.
• For more information and to download the general picture recognition software go to the download page link above.

• Dedicated applications, this software is general picture recognition, it is not dedicated picture recognition.
• Therefore it is not suitable as say face recognition software so it could not identify if two faces were the same two faces but in different pictures. It may recognise with the help of AI that there are two faces that are similar in both pictures but it is not detailed enough to recognise if those two faces were the same two faces in both pictures. This is because it does not use measurements like in face recognition software. However if the faces were close up the same size and both in the middle of the picture it will find both faces without any problem even if they have different backgrounds.
• Its not dedicated therefore is not designed to measure distances between certain similar points.
• Its not any form of OCR therefore could not be used for things like number plate recognition. It does not convert from graphical text to normal asc text. However its ideal for finding similar forms of the same type.
• It does not look for very fine detail at expected locations again moving it away from any sort of dedicated picture recognition. For instance it does not try to measure distance between eyes again moving it away form face recognition.
• This software only runs on PC or Laptops it cannot be used on Mobile devices directly etc., however if your mobile device is connected to your computer and is seen as a normal drive then it can search that device.
• It only searches for pictures than can be searched for via a computer.
• Cannot search for pictures on the Internet. If you download pictures to your harddrive it can search for those pictures but it cannot search for pictures directly off the Internet.

Research and Universities

For those who are interested in picture recognition this software is a good starting point to understanding how picture recognition actually works. I have included some visible parameters if you turn on View Working that really gives a good indication on how the software makes its decisions.

You can turn on View Working and the software will start to give you its reasoning why it has included a picture for you to view or excluded that picture. View working may slow down the recognition time a little but not much but will give the user a good indication of what it is doing.

Eventually all objects will be recognised by computers and will make real sense to the computer program in control. My theory of relative in object idea will eventually allow computer systems to understand and make sense of all objects. This theory will eventually become true reality, this theory is my theory and my theory alone. Version 3 of this software starts this theory of but only with the main object in the picture in itself a very large jump however in the future "relative in object theory" will be the way forward and all roads of thinking will eventually have to comply with this theory.

ABOVE EVERYTHING ELSE THIS SOFTWARE WILL FIND PICTURES THAT YOU NEVER KNEW EXISTED ON YOUR HARD DRIVE.

If you wish to ask me any questions not covered above please send me a question via E-mail gprsg@hotmail.co.uk or click below.