Multi modal Information Processing Systems (MiPS) Lab


The Multi modal Information Processing Systems Lab was established in 2008 at the Department of Electrical Engineering with grants from IIT Kanpur and other funding agencies.


The focus of the Lab infrastructure is to create intelligent environments in which humans and machines (i.e., equipments, rooms, and structure) collaborate with each other in a seamless manner and to deliver the information to the user desiring such an information in linguistic and visual modalities. Hence a transformation of the signal to higher-level semantic information for interpretation is required. For example, a meeting room, an entire building, can be converted to an intelligent environment. To incorporate the functionalities described above, development of the necessary theories, algorithms, and implementation prototypes is quintessential and is the goal of research at MiPS Lab. The primary focus of research currently is on multi-modal (speech, audio, video and text) processing.

Infrastructure :

The Lab has a multi channel audio-visual data acquisition test bed which can record upto 32 channels of synchronized multi-modal data. The test bed can also acquire additional modalities like handwriting, haptic inputs. Figure 1, illustrates a photograph of the test bed being used as a meeting room. Figure 2, gives a schematic of the facility. The test bed is also equipped with high fidelity electret microphones and high resolution cameras capable of distant speech, audio and video signal acquisition. Figure 4, illustrates the computing infrastructure at MiPS Lab. It is equipped with the state of the art data server and three industry grade voice servers based on the open source asterisk software. The voice servers are connected to both analog and digital (ISDN-PRI ) lines capable of handling multiple simultaneous audio and video calls. Figure 3 illustrates an intelligent video retrieval application hosted on this infrastructure.


Department of Electrical Engineering,
IIT Kanpur


Prof.Rajesh Hegde

This email address is being protected from spambots. You need JavaScript enabled to view it.

Fig: 1

Fig: 2