Ankita Jain

Recording electromyography (EMG) data from the upper limb plays a huge role in sensorimotor investigations. Such investigations may require a repeated measures design wherein the experimenter has to place surface electrodes on locations matched to that of the previous sessions. However, this is prone to human error, and the resulting locations decided by the experimenter across sessions might not be precisely matched. This might lead to differences in the corresponding cortical correlates, the estimate of which is not yet known. No tool currently exists that can assist the experimenter in the above problem. We propose a novel tool, Myolocate, that will use a webcam to identify the electrodes placed in the human arm and their corresponding locations with respect to well-known anatomical points in the arm in the first session of the experiment and guide the experimenter, in real-time, to quickly identify the electrode locations in subsequent sessions without having to go through the measurement of anatomical reference to resolve the electrode positions of the previous session. The proposed tool will ease locating electrode position and optimize the data acquisition process.

Sachin Pawar, Girish Palshikar, Ankita Jain, Mahesh Singh, Mahesh Rangarajan, Aman Agarwal, Vishal Kumar, Karan Singh. Proceedings of the The 5th Workshop on Narrative Understanding. 2023.

Fully polarimetric Synthetic Aperture Radar (SAR) data provides the huge amount of information which is useful for classifying earth's surface information. Many techniques are available in the literature for land cover classification. Still, it is a very challenging task to identify various land cover classes. Therefore, in the paper, based on the polarimetric indexes, fully polarimetric SAR data has been used for the identification of various land cover types. Thus, a decision-tree-based approach has been proposed for land cover classification in this paper. The proposed algorithm of decision tree classification scheme is successfully implemented and tested on RADASAT-2 data and validated on another set of RADARSAT-2 data.

Development of a practical stereo vision sensor for real-world applications must account for the variability of high-volume production processes and the impact of unknown environmental conditions during its operation. One critical factor of stereo depth estimation performance is the relative alignment of the stereo camera pair. While imperfectly aligned stereo cameras may be rectified in the image domain, there are some errors introduced by both the calibration recovery and image rectification processes. Finally, additional uncalibrated misalignments, for example due to thermal or mechanical deformation in a harsh automotive environment, may occur which will further deteriorate stereo depth estimation. This paper describes an experimental framework for determining these limits using image processing algorithms, operating on graphically synthesized imagery, with performance envelope validation on real stereo image data.