Speech Recognition for Sindhi Language

Project Title

Speech Recognition for Sindhi Language

Project Area of Specialization Artificial IntelligenceProject Summary

Speech recognition is a process of identifying what a person speaks on a mic or any other similar hardware that detects speech and reflects its meaning into text.

Google, Siri and Python and other speech Speech Recognizer facilitate the users to speak and recognize their words but these recognizer’s write the speech of users in English / English-Roman using with/without Internet.

Our Speech Recognition for Sindhi Language is a cross-platform(Android application and desktop). In addition to that, our speech recognition will work offline speech too.

Project Objectives

We want to develop a Speech Recognition Software which supports our native language Sindhi to facilitate people of Sindh and other regions of Pakistan who communicate in Sindhi Language. And so far no one has ever worked in Sindhi language speech recognition. I initiated Sindhi Speech Recognition for the 1st time. This speech recognition can be integrated in Android Applications which is most commonly used source of interaction. With that it doesn’t require internet connection, as it is offline. It will also be integrated in MS Office.

Project Implementation Method

1.First we will have to input the speech from mic.

2.Then we will extract features from it.

3.Now we will do Decoding with the help of Phonetic dictionary and            Acoustic models

4.Then the outcome of decoding will give us word's sequence, from which we will then search words from it using language model.

5.After searching, we will have our final output in the form of Text/Sentence.

The process works in a way that we have to record some voices of words which we want to train via mic, then we have to create written words of those recordings. After that we need to link that words with voices/recordings in order to create key values. Here key would be the voice, and value would be the word, which will be printed.

Next time when someone says the same word on mic, the frequency and wave lenght with amplitude of the spoken word will be matched with the previous recorded words. If the spoken word has resemblance with recorded words, the stored value of word will be printed.

In this way the speech recognition words and same procedure we are following.

Benefits of the Project

There are so many benefits of this project.

1. Conversing is a major mode of communication and recognition of that words are important to operate and make machines learn.

2. Users of Sindhi language can easily just speak the words and their words would be printed on the screen, which they can forward via sms or any other means of communication.

3. Users can controls their Pc/Mobile via just speaking Sindhi Language words.

4. Users can have advantage that this system can be integrated with MS office, they can simply open MS Word, start speaking and all their spoken words will be typed in MS Word.Documents can be generated up to three times as fast with speech recognition as they can if they are typed

5.Work processes become more efficient because document processing times become shorter.

6. Speech recognition can play a great role in Dictionary where people do not know exact spelling, by just speaking they can have the accurate word with its means. Saves alot of time.

There are enormous uses of speech recognition specially in native language.

Technical Details of Final Deliverable

The Final Deliverable will contain:

1. A software compatible with Windows/Mac Operating System which can control many functions like switching Off Pc, Copying files, Sending Emails like stuff and many more. A fully automated sindhi speech recognized laptop/pc.

2. An Add-in for MS Word via which MS Word can recognize Sindhi Langauge and type the words of Sindhi Language

3. An Android Application Software compatible with Android Operating System which can control functions on speech like Making a call, Sending sms, Opening YouTube for playing song, setting up alarms and remainders and so on.

4. A keyboard of Sindhi Language which can also pronounce the words written.

Final Deliverable of the Project Software SystemType of Industry Education , IT Technologies Artificial Intelligence(AI)Sustainable Development Goals Good Health and Well-Being for People, Quality Education, Decent Work and Economic GrowthRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	38178
Deity V-Mic D3 Pro Super-Cardioid Directional Shotgun Microphone	Equipment	1	28178	28178
1 TB External HDD with case	Equipment	1	10000	10000