0

Hi everyone,

I am doing my final year in BSc Computing and Information systems.
I am looking for some advice on my chosen project.

Its about a mobile based sign translator.Photos that contain some text in foreign language taken using mobile camera is sent to a work station where the text is detected from the image and recognised perhaps by using OCR techniques and then translated to target language(for example: french to english or chinese to english).The translation is sent back to mobile phone.

I have 8 months to complete my project and I would like to know if it is possible to start and finish this project within 8 months.
Using OCR techniques to detect text from iimages has been done in the past as a separate project which could hv taken some reasonable time.
Also the translation of one language to another has been done as a separate project in the past.

So my doubt is whether all these can be integrated to finish the project in 8 months??? Or may be if we can use existing OCR software in the project if its allowed...then it will make things easy.But is it acceptable to do that?The question arises then wat work is to be done by us if we r to use existing software in the project(is it just using all the software with a mobile)

Please please please.......... provide me with some advice and tips regarding this as to what is supposed to be done and which parts of the project can be narrowed through some other means instaed of doing all by ourselves.

Please help me.....!!!!!! :'(

10
Contributors
25
Replies
33
Views
10 Years
Discussion Span
Last Post by Marseline
0

Hi,

Thanks for replying to my thread.Mine is an individual project.

I have got another project topic as well which is called automated image annotation similar to that of google image search but the image tagging will be done automatically to ease the process of keyword-based search.

Appreciate ur advice on both these projects.Pls reply soon :)

Many thanks !!!!

how many team members?

0

I see they are great and I recommend to take Google as sponsor to your idea, in case you want to sell your project or work in Google

0

Hi Ramy,

Thanks for your response.I was really scared wat u wud say to my ideas.But am happy tht it turned out to be a positive response.My primary fear in proceeding with these projects was tht its an individual project and I have 7 months to complete it along with other courseworks for 6 other subjects.

So my doubt was if it is manageable to complete it in 7 months time.....or if necessary whether I cud narrow the scope down further.If u think the projects are toooo broad, pls i wud appreciate any suggestions on ur behalf on how best to narrow the scope for each project.

Pls consider the mobile based sign translator project as mentioned in my 1st thread.

Sure I will think abt takin Google as sponsor....in fact its wat my supervisor adviced too...it depends on the possibility frm my country as well(am frm sri lanka studyin Univeristy of London external programme)

Waitin for ur reply .

Thanks a lot again.Your advice has really been invaluable among all others tht I got.It has given me real confidence

Thanks a lot for it again :):):)

Vashini

0

First of all, Vashini, you don't need to develop complete project, any project and especially GP. there is something called "Further Work" it's the work to be done after your submission, almost of us develop 80% of the project, how to determine that, just select any idea, try to design it, from your design (Logic design) you'll estimate the time need. plus your experience in the tool used and the programming language in supporting for such logic need i.e MATLAB better than C# in huge math. calculations or C# better than VC++ in ADO application.
and if you use .net for example and you have a lot of components you can use the appropriate language for each component and integrate between them later anyway that's not our issue, what I want to say it select good idea for 7 months DON'T select idea could be done in 3 months (that's would be negative!!).
OCR, is very good and you can use it say in traffic monitoring, get the car plate number and car model that exceeds specific speed another application you can use OCR, you can develop automatic garage, once a car get it, you count the time it was in, calculate the fees before getting out you enforce the driver to pay the fees, OCR is very large and very fantastic, use it but try to search in the business side.
Hope to you the best
sorry for long reply :)

0

Hi Ramy,

I don't think ur message is long compared to mine anyway.....if anyone is to be sorry am more sorry than u for my long messages.

I had the idea of doin plate number recognition system for my project but then i dropped it coz my supervisor was not happy abt it :(

Later i came across the idea of automated image recognition......so before taking this idea to my supervisor I just want to knw if this idea is very complex to be done as an individual project.

If you can suggest me anything(any ideas u may have abt how it can be done) relating to this ...... so tht I can use it as a basis start, it wud be really great !!!!

Sorry for my long msg again.......nd sorry for continuing this thread further....I just need some more clarification for the AUTOMATED IMAGE ANNOTATION project...

Waiting for ur response :)


First of all, Vashini, you don't need to develop complete project, any project and especially GP. there is something called "Further Work" it's the work to be done after your submission, almost of us develop 80% of the project, how to determine that, just select any idea, try to design it, from your design (Logic design) you'll estimate the time need. plus your experience in the tool used and the programming language in supporting for such logic need i.e MATLAB better than C# in huge math. calculations or C# better than VC++ in ADO application.
and if you use .net for example and you have a lot of components you can use the appropriate language for each component and integrate between them later anyway that's not our issue, what I want to say it select good idea for 7 months DON'T select idea could be done in 3 months (that's would be negative!!).
OCR, is very good and you can use it say in traffic monitoring, get the car plate number and car model that exists specific speed another application you can use OCR, you can develop automatic garage, once a car get it, you count the time it was in, calculate the fees before getting out you enforce the driver to pay the fees, OCR is very large and very fantastic, use it but try to search in the business side.
Hope to you the best
sorry for long reply :)

0

"I had the idea of doin plate number recognition system for my project but then i dropped it coz my supervisor was not happy abt it"
my advice is to give him good proposal, which contains good abstract about your Idea, your customers (which people make use of it -virtually-) and your passion to make him/her live the idea with you and give you the support.

For your "automated image recognition" idea give me scenario? How the user will use it?

I'll think but give me your interested fields (Software Engineering, Image Proccessing, AI, ....)
you don't need to be sorry man, just ask :)

0

Hi Ramy,

Yep sorry.I dint mention abt the scenario. My idea is to narrow the process of assigning automatic keywords to unannotated medical images.Yes...its to do with medical images.

In a given situation where a doctor mite need to refer similar cases of past patients for the diagnosis.In tht case he may need to retrieve similar images using a keyword....such as "Heart". When he searches through this keyword, the system shud return all heart images in the database.

Since the topic is called automatic image annotation........we can just try to solve the problem of assigning approriate keywords to unassigned medical images in the database.

This is the basic scenario.But i dont knw if this information helps u to consider abt it.

Sorry for the long message again.AI is also a field tht am interested in.Am OK with Image processing too.

Waiting for ur response.Many thanks !!! :)

"I had the idea of doin plate number recognition system for my project but then i dropped it coz my supervisor was not happy abt it"
my advice is to give him good proposal, which contains good abstract about your Idea, your customers (which people make use of it -virtually-) and your passion to make him/her live the idea with you and give you the support.

For your "automated image recognition" idea give me scenario? How the user will use it?

I'll think but give me your interested fields (Software Engineering, Image Proccessing, AI, ....)
you don't need to be sorry man, just ask :)

0

so, you won't save any information about the image in the database!!
just a huge number of images when I want to get "Heart" images, you will use AI algorithms to get them.
or you want before pushing the images into DB, you use AI algorithms to identify it and save image and its keyword for optimizing search?

I asked about that, because your core work would be in this.

you didn't comment about to give him a proposal about OCR!!

0

"just a huge number of images when I want to get "Heart" images, you will use AI algorithms to get them."

This is referred as CBIR(Content based image retrieval).we will give the system a sample image of Heart and then the system will retrieve features from the sample image and compare those features with the images in the database.The closely matched ones will be given as results to the search for Heart

"or you want before pushing the images into DB, you use AI algorithms to identify it and save image and its keyword for optimizing search?"

This one sounds like Automated Image annotation tht I wanted to do for my project.For example : there will be lot of medical images and their related descriptions stored in the database.Some images mite not have any descriptions in the DB.So when a user searches for a Heart image using the keyword "heart"........the Heart images in the DB tht dont have any description like "Heart" wont be given as results to user.In this case ....the system shud learn to analyse the contents of each image in the DB and assign appropriate keywords to them.....in this way it will optimize search.To do this ....we can use some training set of images and their keywords to train the system to assign closely related keywords to the images tht dont have any descriptions.....

Hope this makes it clear......for u to get a feeling of wat it is further u can refer this web site....http://www.alipr.com :)

so, you won't save any information about the image in the database!!
just a huge number of images when I want to get "Heart" images, you will use AI algorithms to get them.
or you want before pushing the images into DB, you use AI algorithms to identify it and save image and its keyword for optimizing search?

I asked about that, because your core work would be in this.

you didn't comment about to give him a proposal about OCR!!

0

"just a huge number of images when I want to get "Heart" images, you will use AI algorithms to get them."

This is referred as CBIR(Content based image retrieval).we will give the system a sample image of Heart and then the system will retrieve features from the sample image and compare those features with the images in the database.The closely matched ones will be given as results to the search for Heart

"or you want before pushing the images into DB, you use AI algorithms to identify it and save image and its keyword for optimizing search?"

This one sounds like Automated Image annotation tht I wanted to do for my project.For example : there will be lot of medical images and their related descriptions stored in the database.Some images mite not have any descriptions in the DB.So when a user searches for a Heart image using the keyword "heart"........the Heart images in the DB tht dont have any description like "Heart" wont be given as results to user.In this case ....the system shud learn to analyse the contents of each image in the DB and assign appropriate keywords to them.....in this way it will optimize search.To do this ....we can use some training set of images and their keywords to train the system to assign closely related keywords to the images tht dont have any descriptions.....

Hope this makes it clear......for u to get a feeling of wat it is further u can refer this web site....http://www.alipr.com :)

so, you won't save any information about the image in the database!!
just a huge number of images when I want to get "Heart" images, you will use AI algorithms to get them.
or you want before pushing the images into DB, you use AI algorithms to identify it and save image and its keyword for optimizing search?

I asked about that, because your core work would be in this.

you didn't comment about to give him a proposal about OCR!!

0

But Ramy....how long wud this project require to complete(i hv only 7 months) and do u think the idea is tooo large for a BSc degree project.

I havn't done any projects b4 and thts wat scares me most .Pls I need ur advice on this.

good idea, go on (Y)

0

7 months is good, just begin right, design your project don't hurry up to write code, TAKE CARE, design design design, this idea is suitable for GP.

0

Thanks Ramy.....Thanks for everything....I wud keep u updated on the status of my project.
Cheers !!!!! :)

Also if u prefer please send ur chat id as a private message to me so tht we can chat sometimes.

Thanks again .Bye bye :):)

7 months is good, just begin right, design your project don't hurry up to write code, TAKE CARE, design design design, this idea is suitable for GP.

0

Hi ,

It is possible.


where are you going to do your project?


Regards,
Raja

Hi everyone,

I am doing my final year in BSc Computing and Information systems.
I am looking for some advice on my chosen project.

Its about a mobile based sign translator.Photos that contain some text in foreign language taken using mobile camera is sent to a work station where the text is detected from the image and recognised perhaps by using OCR techniques and then translated to target language(for example: french to english or chinese to english).The translation is sent back to mobile phone.

I have 8 months to complete my project and I would like to know if it is possible to start and finish this project within 8 months.
Using OCR techniques to detect text from iimages has been done in the past as a separate project which could hv taken some reasonable time.
Also the translation of one language to another has been done as a separate project in the past.

So my doubt is whether all these can be integrated to finish the project in 8 months??? Or may be if we can use existing OCR software in the project if its allowed...then it will make things easy.But is it acceptable to do that?The question arises then wat work is to be done by us if we r to use existing software in the project(is it just using all the software with a mobile)

Please please please.......... provide me with some advice and tips regarding this as to what is supposed to be done and which parts of the project can be narrowed through some other means instaed of doing all by ourselves.

Please help me.....!!!!!! :'(

-1

Hi
r u good in ai.if yes ,y u would like go for image procesing which is somewat tedious...waiting for ur reply...but i like d way u expressed ur idea n thoughts.

Votes + Comments
Another pointless "project" thread bumped by some illiterate script kiddie :(
-1

hi everyone
Now I am in the way doing final year project. But I want help for choosing/selecting what type of project should I do?
I am medium in computer programing. So pls help me in what area should I do my project
thank U!!!!!!!!!!!!!

Votes + Comments
Did you read the rest of the thread?
-1

hi i ansari sabiya . m a student of TY comp sci.....i m working on a project regarding the constructions management. can u please sugest me a sutable topic for my project or the which i have selected is correct. and i also need some informations regarding the same for my synopsis which have to b submitted by me before 30th june. So please help me out with it as soon as possible.

thank you-

0

OK assuming yours is text based image processing, based on some vague text color input, you could isolate the part with filters. MATLAB can help extract the data as Ramy already suggested. If you are more familiar with MATLAB, your work should be much faster and easier... look up several webinars from their site for image processing help.

Try running some sort of a skeletonizing algo to extract the bones of the individual characters. In each language, you will notice that the bone-map of characters of the alphabet differ. This is obviously if you did not get permission to use direct OCR.

Hey... i can see you do it in 7 months! Just try and not worry so much.

-1

Hi everyone,

I am doing my final year in BSc Computing and Information systems.
I am looking for some advice on my chosen project.

Its about a mobile based sign translator.Photos that contain some text in foreign language taken using mobile camera is sent to a work station where the text is detected from the image and recognised perhaps by using OCR techniques and then translated to target language(for example: french to english or chinese to english).The translation is sent back to mobile phone.

I have 8 months to complete my project and I would like to know if it is possible to start and finish this project within 8 months.
Using OCR techniques to detect text from iimages has been done in the past as a separate project which could hv taken some reasonable time.
Also the translation of one language to another has been done as a separate project in the past.

So my doubt is whether all these can be integrated to finish the project in 8 months??? Or may be if we can use existing OCR software in the project if its allowed...then it will make things easy.But is it acceptable to do that?The question arises then wat work is to be done by us if we r to use existing software in the project(is it just using all the software with a mobile)

Please please please.......... provide me with some advice and tips regarding this as to what is supposed to be done and which parts of the project can be narrowed through some other means instaed of doing all by ourselves.

Please help me.....!!!!!! :'(

Hi how did you come up with your project idea? Do you maybe still have ideas for me i got less then 6 months so its should not be time consuming
thanks in advance

This topic has been dead for over six months. Start a new discussion instead.
Have something to contribute to this discussion? Please be thoughtful, detailed and courteous, and be sure to adhere to our posting rules.