Automatic Robotic-Assisted Diffuse Reflectance Spectroscopy Scanning System
Kaizhong Deng, Christopher J. Peters, George P. Mylonas, Daniel S. Elson · 2025
Read Automatic Robotic-Assisted Diffuse Reflectance Spectroscopy Scanning System by Kaizhong Deng, Christopher J. Peters, George P. Mylonas, Daniel S. Elson (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-08470.txt.
Introduction
Kaizhong Deng, Christopher Peters, George Mylonas, and Daniel Elson automate light based tissue scans with a robot, in work from 2025. Diffuse reflectance spectroscopy reads tissue makeup and can flag tumour left behind in surgery. Hand held point checks sample small spots, so wide area coverage stays slow and uneven across operators.
A robot arm scans large areas on its own with mixed vision control for guidance. A height module holds probe contact steady as the surface shape shifts during the sweep. Tests showed scan paths matched commands and spectra matched hand gathered results closely. Hand collection is the current gold standard for this tissue check in clinics. The team sees theatre use that frees the surgeon from manual sweeps.
Catalog record
| CREATOR | Kaizhong Deng, Christopher J. Peters, George P. Mylonas, Daniel S. Elson |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 30 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-08470
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Automatic Robotic-Assisted Diffuse Reflectance Spectroscopy Scanning System Kaizhong Deng, Christopher J. Peters, George P. Mylonas, Daniel S. Elson Abstract Diffuse Reflectance Spectroscopy (DRS) is a well-established optical technique for tissue composition assessment which has been clinically evaluated for tumour detection to ensure the complete removal of cancerous tissue. While point-wise assessment has many potential applications, incorporating automated large-area scanning would enable holistic tissue sampling with higher consistency. We propose a robotic system to facilitate autonomous DRS scanning with hybrid visual servoing control. A specially designed height compensation module enables precise contact condition control. The evaluation results show that the system can accurately execute the scanning command and acquire consistent DRS spectra with comparable results to the manual collection, which is the current gold standard protocol. Integrating the proposed system into surgery lays the groundwork for autonomous intra-operative DRS tissue assessment with high reliability and repeatability. This could reduce the need for manual scanning by the surgeon while ensuring complete tumor removal in clinical practice. I INTRODUCTION The full resection of a tumour, while preserving healthy tissue, is one of the most critical objectives for many types of cancer surgery, which is correlated to minimal local recurrence and prolonged disease-free survival [ 1 ] . To ensure optimal results, it is necessary to guarantee that no tumour is present at the edges of the removed specimen. Thus, there is a demand for intra-operative tissue assessment tools capable of characterising tissue in real time. Several optical sensing technologies have been shown to facilitate real-time, non-invasive tissue classification [ 2 , 3 , 4 ] . Diffuse Reflectance Spectroscopy (DRS) is one of the most well-established techniques for tissue assessment which has recently been shown to be accurate at distinguishing between normal and cancer in upper [ 2 ] and lower [ 5 ] gastrointestinal tissues. It has the advantages of low cost, high accuracy, ease of use and real-time signal analysis. DRS operates by illuminating the tissue with a band of light and measures the reflected light after its interaction with the tissue through scattering and absorption [ 6 ] . The analysis of the spectral response allows identification of the optical properties of the tissue, hence determining its composition. DRS is usually configured as a hand-held probe, allowing surgeons to manually position it at a target location and requires steady tissue contact to obtain high quality data. Fig. 1: Overview of the automatic DRS sampling system : previous DRS sampling relies on manual operation to position the DRS probe. This proposed system utilised the visual servoing module and contact height compensator to the robot to automatically position the probe. Two cameras capture views from different scales as feedback to the system. Fig. 2: DRS automatic system : The left top is a zoomed view of a third-person camera image showing two image features: probe tip and illuminated area centre (also named as light centre), and a target point represented by a marker. The bottom left image illustrates the two stages of a DRS sampling procedure: the approach stage and the scanning stage. The right section is a diagram of the overall control. The system takes images from both cameras to extract the current visual state using two pairs of features and probe contact height. The visual servoing generates the motion to execute the scanning trajectory while tracking the target contact height. The resulting action drives the robot to sample DRS on the tissue surface. Despite the success for single-site sampling, the utility of large-area in vivo sampling has not yet been realised. A significant challenge of manual DRS collection is its high variability due to unsteady hand movement and variable contact pressure from the operator [ 7 , 8 ] . For example, the contact pressure exhibits a non-linear relationship to the acquired spectrum [ 9 ] , which can cause different spectral responses for the same tissue type. This is also the main reason for the intra-observer variance between different operators, leading to unreliable results. An automated robot-assisted scanning system capable of controlling the contact condition could enhance the consistency of DRS in highly structured large-area tissue scanning scenarios. Such a system could standardise the DRS scanning procedure, generate highly standardised and trustworthy results, and free practitioners from tedious manual scanning. Previous research on robotic-assisted spectroscopy primarily focused on ex vivo specimens. For example, some robotic scanning systems require specimens to be prepared as microscope slides [ 10 ] or cut into smaller pieces [ 11 ] . However, these systems require meticulous specimen preparation, including dissection into slides and smaller samples, which is not feasible in a real-time surgical setting, making such methods impractical for use during surgery. Thus, in vivo tissue scanning should accommodate to its original shape or size. Recently, robotic systems have been developed for DRS scanning of large meat samples [ 12 ] and organ phantoms [ 13 ] . However, these systems lack contact control to maintain constant contact, which might affect the results. Visual servoing has been widely used to control robotic systems using images as feedback [ 14 ] and has been applied to medical robotic systems [ 15 ] . For instance, it has enabled surgical robots to automatically position a laser instrument holder [ 16 ] , control the scanning of microscopic probes [ 17 ] , and control the endoscope [ 18 ] . Additionally, it has been used to control an instrument mounted on a robot arm for automatic middle ear surgery [ 19 ] and for generating pCLE mosaics [ 20 ] . Thus, visual servoing is well-suited to position the DRS probe precisely with a robotic system. This study proposes a framework for a robotic-assisted autonomous DRS scanning system. It utilised a hybrid visual servoing approach with a precise height compensation module to acquire a consistent DRS spectrum as shown in Fig. 1 . To the best of our knowledge, this is the first achievement of automatic robotic scanning using DRS with a contact condition management and without requiring any further tissue preparation steps. The primary contributions of this study are as follows: • Integration of the DRS instrument into a robotic system, controlled via visual servoing. • Development of a precise height estimation and compensation module based on a deep neural network for real-time feedback. • Evaluation of the system on two phantom tissue models and two animal tissue samples, demonstrating its trajectory precision and spectrum consistency to be comparable to manual collection, which is the current gold standard collection protocol. II PROBLEM STATEMENT Task description. The sampling of DRS on a target surface consists of two stages: the approach stage and the in-contact scanning stage. During the approach stage, the DRS probe is moved to the start of a planned path on tissue. Once the firm contact has been established, the robot should switch to the in-contact scanning stage to automatically scan along a path indicated by the surgeon on the monitor of a third-person view camera image. The sampling should be made at the standard contact plane where the minimal pressure is made. The DRS spectrum collected along the scanned path is recorded during the in-contact scanning stage. Spectrum evaluation. The reflectance spectrum