ISHAAN: A System for Bidirectional Machine Translation Between 1) English and Assamese, Bodo, Manipuri, Nepali 2) Manipuri and Hindi 3) Assamese and Bodo

Department:	Department of Computer Science and Engineering
Sanctioned by:	Ministry of Electronics and Information Technology (MeitY), Govt. of India
Sanction number:	11(14)/2018-HCC(TDIL)
Amount:	INR 2.44 Crores
Area of Research:	Natural Language Processing
Department:	Department of Computer Science and Engineering
Date of commencement:	01 April, 2022
Date of completion:	31 March, 2025
PI:	Dr. Nongmeikapam Kishorjit Singh, Dept. of CSE, Indian Institute of Information Technology Manipur, Imphal India.
Chief-PI:	Prof. Pushpak Bhattacharyya, Dept. of CSE, Indian Institute of Technology Bombay, Maharashtra, India.

Multi-domain text-to-text machine translation system in synchrony with speech consortia to facilitate bidirectional translation.

Broad Objectives:

1. To create and nurture an ecosystem involving start-ups, Central/State government agencies, working together to develop and deploy innovative products and services in North-East Indian languages.
2. Parallel Text Corpus creation: Multi-words (MW), Named Entity (NE) and Part-of-speech (POS) tagged corpus of size approximately 25K sentences for each of these languages pairs.
3. Development of Machine Translation of each pair.
4. Development of benchmark data system for Machine Translation of each pair. Development of standards for evaluation and corpora creation.
5. Develop Machine Translation APIs which can be integrated into different language-solution-related applications.
6. Develop Mobile-app, Web-service which can be used to access the MT Development of benchmark data system.
7. Deploy the complete Machine Translation software to be accessible to the common mass.

Status: Project currently ongoing.

Prototype Development of Self-biased High-Power Circulator for Particle Accelerators

Department:	Department of Electronics and Communication Engineering
Sanctioned by:	Board of Research in Nuclear Sciences (BRNS), Department of Atomic Energy (DAE), Govt. of India.
Sanction number:	59/14/09/2019-BRNS
Amount:	INR 33.13 Lakhs
Date of commencement:	01-April-2020
Date of completion:	31-March-2023
PI:	Dr Subasit Borah, Dept. of ECE, Indian Institute of Information Technology Manipur, Imphal India.
PC:	Shri Prashant Pareek, Pulsed magnet & Ferrite Technology Lab, Accelerator Magnet Technology Division, Raja Ramanna Centre for Advanced Technology (RRCAT) Indore.
Co-PC:	S. Senthil Kumar, SO-F, Pulsed magnet & Ferrite Technology Lab, Accelerator Magnet Technology Division, Raja Ramanna Centre for Advanced Technology (RRCAT) Indore.

Objectives:

1. To design and simulate compact and high performance self-biased circulators both at low and high-power levels
2. Fabrication of the highly anisotropic ferrites for temperature stable high-power circulator
3. Characterization of high anisotropic ferrite and ferrite metamaterials like FMR, spin wave resonance, tensor permeability, non-reciprocal propagation in nano sized ferrite resonators at the PC’s laboratory and PI will also be involved for testing and analysis of the results.
4. Fabrication of the designed prototype of the self-biased circulator and its experimental validation at low power level of microwaves at PC’s laboratory
5. Final design transfer from the PI’s laboratory to the end user, i.e., PC’s laboratory for testing the device under practical environments.

Studies carried out:

1. Design and simulation of circulators in stripline and waveguide configurations with standard ferrites has been done. This has been helpful in establishing the design procedure for self-biased circulators.
2. Studied calculation of permeability tensor in highly oriented hexaferrites. Permeability tensor is a fundamental concept for finding a circulation adjustment. Knowing the internal DC magnetic field and other ferrite characteristics such as saturation magnetization, anisotropy field, and linewidth, allows for its calculation.
3. It turns out that the internal DC magnetic field arising out of the high anisotropy field of the hexaferrite materials available today is too high for our target frequency range.
4. Test simulations were performed using fictitious hexaferrite materials with reduced anisotropy field, which suggests that if those materials are synthesised, circulators may be made utilising them.

Status: Completed

A Multilingual Reference WordNet on Four of the Major Indigenous Dialects of Manipur and Manipuri Language

Department:	Department of Computer Science and Engineering
Sanctioned by:	Science and Engineering Research Board (SERB), Dept. of Science and Technology (DST), Govt. of India
Sanction number:	CRG/2019/006887
Amount:	INR 41.81 Lakhs
Date of commencement:	10 February, 2020
Date of completion:	09 February, 2023
PI:	Dr. Nongmeikapam Kishorjit Singh, Dept. of CSE, Indian Institute of Information Technology Manipur, Imphal India.
Co-PI:	Dr. Kabita Thaoroijam, Dept. of CSE, Indian Institute of Information Technology Manipur, Imphal India.
Co-PI:	Dr. Leihaorabam Sarbajit Singh, Dept. of HBS, Indian Institute of Information Technology Manipur, Imphal India.

Objectives:

1. Creation of a Corpus by collecting lexical items (i.e., words or phrases) of different dialects.
2. Creation of WordNet for each of the dialects.
3. Creation of a single open online software to host the WordNet.
4. Thousands of synonyms sets or synsets to be created.

Concise Research accomplishment:

The project yielded 6 monolingual unique word corpus for Manipuri Bengali, Manipuri Meetei-Mayek, Tangkhul, Thadou, Mao and Kabui/Rongmei. Manipuri Bengali and Manipuri Meetei-Mayek corpus consisted of 53940 words, Tangkhul consisted of 21651 words, Thadou of 22247 words, Mao of 41402 words and Kabui/Rongmei of 40927 words. The project also yielded 21 parallel corpora involving the combination of 7 monolingual sentences viz. Manipuri (Bengali), Manipuri (Meetei-Mayek), Tangkhul, Thadou, Mao, Kabui/Rongmei and English. The project created 6 WordNets, with detailed lexical data of 52548 synsets and 53940 related words for Manipuri Bengali and Manipuri Meetei-Mayek WordNet, 20973 synsets and 21651 related words for Tangkhul WordNet, 21491 synsets and 22247 related words for Thadou WordNet, 38547 synsets and 41402 related words for Mao WordNet and 38833 synsets and 40927 related words for Kabui/Rongmei WordNet. These wordnets are connected to make it a multilingual WordNet for Manipuri Bengali, Manipuri Meetei Mayek, Tangkhul, Thadou, Mao and Kabui/Rongmei. The project developed a mass accessible website at www.kangleiwordnet.in, where the multilingual wordnet and other resources are hosted and made available to use by the mass. The project also helped develop the first efficient mass usable Meetei Mayek Keyboard Input Method Editor (IME).

Status: Completed

A prototype of Auto Navigating E-cart using multispectral imaging

Department:	Department of Computer Science and Engineering
Sanctioned by:	Ministry of Electronics and Information Technology (MeitY), Govt. of India
Sanction number:	4(6)/2018-ITEA and 18/02/2019
Amount:	INR 40.00 Lakhs
Date of commencement:	18 February, 2019
Date of completion:	10 October, 2022
PI:	Dr. Nongmeikapam Kishorjit Singh, Dept. of CSE, Indian Institute of Information Technology Manipur, Imphal India.

Broad objectives

1. Image Classification using SLIC Superpixel and FAAGKFCM Image Segmentation
2. Image and Ultrasonic Sensor Fusion for Object Size Detection
3. Stereo System based Distance Calculation of an Object in Image
4. Development of image fusion algorithm using RGB- NIR images.
5. Development of parametric semantic segmentation.
6. Development of multiple object detection algorithm.
7. Development of an integrated non-parametric and parametric scene parsing algorithm.
8. Modelling a replica of an autonomous RC toy Car.
9. Development of image stitching and blending using SURF.
10. Modelling a replica of the e-cart in an RC toy car.
11. Lane detection algorithm for non-lane based Unstructured Road.
12. Path finding using GPS data.

Concise Research accomplishment:

The focus target for development of navigation, planning & control algorithm achieved. Also, the integration of sensor with the E-cart and to start real time application of the algorithms design done. In concise, the following were achieved:

Object detection

Distance Estimation: Monocular distance estimation, Stereo vision distance estimation

Auto Navigation: Source to Destination auto navigation, Semantic Segmentation based Path planning

Self Driving three Subsystem: Forward/ Reverse movement Subsystem, Braking Subsystem, Steering Subsystem

Status: Completed