Publications

RUSOpt: Robotic UltraSound Probe Normalization with Bayesian Optimization for In-Plane and Out-Plane Scanning 

IEEE International Conference for Automation Science and Engineering (CASE) 2023

Authors: Deepak Raina, Abhishek Raman Mathur, Richard M Voyles, Juan Wachs, Sh Chandrashekhara, Subir Kumar Saha 

Abstract

Abstract

Computer-assisted interventions for enhanced experience in minimally invasive surgery (MIS) require motion tracking of the surgical instruments. Instrument tracking in conventional and robot-assisted MIS is a challenging problem due to the availability of only 2D projections from the camera and the possibility of minimal hardware integration. This paper aims to track and visualize the movement of the entire instrument (both shaft and the metallic clasper) to intervene within the surgical environment with safe navigation. Segmentation of surgical instruments plays a vital role in tracking the instrument. Unlike existing works, we develop a 2D tracker based on the segmentation maps of surgical instruments in terms of intervals to enable preparations of a labeled data-set having no or minimal knowledge of ground truth. The motion is estimated by expressing the geometric changes in the 2D intervals. The results are processed using algorithms based on kinematics to generate 3D tracking information. Synthesized and experimental results in 2D and 3D motion estimates show that errors are negligible for the method to be used for labelling and motion tracking of instrument for various applications. The results conclude that utilizing the proposed simple technique of 2D segmented intervals and their geometrical changes for 3D tracking is a promising computationally low, direct plug-in to regular MIS practices for 3D visualization of the surgical instrument.

One of the significant challenges faced by autonomous robotic ultrasound systems is acquiring high-quality images across different patients. The proper orientation of the robotized probe plays a crucial role in governing the quality of ultrasound images. To address this challenge, we propose a sample-efficient method to automatically adjust the orientation of the ultrasound probe normal to the point of contact on the scanning surface, thereby improving the acoustic coupling of the probe and resulting image quality. Our method utilizes Bayesian Optimization (BO) based search on the scanning surface to efficiently search for the normalized probe orientation without using the ultrasound image data. We formulate a novel objective function for BO that leverages the contact force measurements and underlying mechanics to identify the normal. We further incorporate a regularization scheme in BO to handle the noisy objective function. The proposed strategy has been validated on urinary bladder phantoms with planar, tilted and curved surfaces with varying search space limits. Further, simulation-based studies have been carried out using 3D human mesh models representing the actual physiology of humans. The results demonstrate that the mean (±SD) absolute angular error averaged over all phantoms and 3D models is 2.4 ± 0.7◦ and 2.1 ± 1.3◦, respectively.

From 2D to 3D Surgical Instrument Tracking: A Technique Based on Intervals and Geometric Cues 

International Journal of Medical Robotics and Computer Assisted Surgery, Wiley

Authors: Shubhangi Nema, Abhishek Raman Mathur, Leena Vachhani 

-> Submitted and Under Review

Abstract

Background

This paper addresses the challenge of instrument tracking in minimally invasive surgery (MIS), crucial for computer-assisted interventions. Conventional and robot-assisted MIS face difficulties due to limited 2D camera projections and minimal hardware integration.

Method

The objective is to track and visualize the entire surgical instrument’s movement, including the shaft and metallic clasper, enabling safe navigation within the surgical environment. The proposed method involves 2D tracking based on segmentation maps, allowing the creation of a labeled dataset without extensive ground truth knowledge. Geometric changes in 2D intervals express motion, and algorithms based on kinematics process results into 3D tracking information.

Result

Synthesized and experimental results in 2D and 3D motion estimates demonstrate negligible errors, validating the method for labeling and motion tracking of instruments in futuristic applications.

Conclusion

The conclusion emphasizes the simplicity and computational efficiency of the proposed 2D segmentation-based technique, highlighting its potential as a direct 

plug-in for 3D visualization in regular MIS practices.

Towards Practical Robot Surgery: Harnessing Joint Analysis with Reinforced Learning for Real-World Application 

Working Paper - For submitting at International Conference on Intelligent Robots and Systems (IROS) & IEEE Robotics and Automation Letters 

(Representing IIT Goa as an undergraduate researcher for this endeavor)

In association with IIT Bombay