Maki Sugimoto’s research while affiliated with Keio University and other places

What is this page?


This page lists works of an author who doesn't have a ResearchGate profile or hasn't added the works to their profile yet. It is automatically generated from public (personal) data to further our legitimate goal of comprehensive and accurate scientific recordkeeping. If you are this author and want this page removed, please let us know.

Publications (227)


Facial Gesture Classification with Few-shot Learning Using Limited Calibration Data from Photo-reflective Sensors on Smart Eyewear
  • Conference Paper

December 2024

·

2 Reads

·

Maki Sugimoto

·

Brian Iwana








Fig. 1 First-person-view surgical video recorded by a wearable camera was downsampled and passed through Mask R-CNN to identify body tissues and surgical tools. RGB surgical images and region masks were
Fig. 2 Example of the results of segmentation. a RGB surgical images. b The result of body tissues segmentation. Fat is painted magenta and muscle is painted green. Dermis is also predicted in the center image,
Fig. 3 Graph illustrates the changes in the sizes of body tissue regions and the workflows. The abbreviations used in the graph are as follows: OB (other behavior), BI (before incision), F (making flap), ML (mam-
Number of labeled images for procedural workflow recognition
IoU and AP of body tissues segmentation

+1

An analysis on the effect of body tissues and surgical tools on workflow recognition in first person surgical videos
  • Article
  • Full-text available

February 2024

·

31 Reads

International Journal of Computer Assisted Radiology and Surgery

Purpose Analysis of operative fields is expected to aid in estimating procedural workflow and evaluating surgeons’ procedural skills by considering the temporal transitions during the progression of the surgery. This study aims to propose an automatic recognition system for the procedural workflow by employing machine learning techniques to identify and distinguish elements in the operative field, including body tissues such as fat, muscle, and dermis, along with surgical tools. Methods We conducted annotations on approximately 908 first-person-view images of breast surgery to facilitate segmentation. The annotated images were used to train a pixel-level classifier based on Mask R-CNN. To assess the impact on procedural workflow recognition, we annotated an additional 43,007 images. The network, structured on the Transformer architecture, was then trained with surgical images incorporating masks for body tissues and surgical tools. Results The instance segmentation of each body tissue in the segmentation phase provided insights into the trend of area transitions for each tissue. Simultaneously, the spatial features of the surgical tools were effectively captured. In regard to the accuracy of procedural workflow recognition, accounting for body tissues led to an average improvement of 3 % over the baseline. Furthermore, the inclusion of surgical tools yielded an additional increase in accuracy by 4 % compared to the baseline. Conclusion In this study, we revealed the contribution of the temporal transition of the body tissues and surgical tools spatial features to recognize procedural workflow in first-person-view surgical videos. Body tissues, especially in open surgery, can be a crucial element. This study suggests that further improvements can be achieved by accurately identifying surgical tools specific to each procedural workflow step.

Download


Citations (55)


... [9, 16,17] [18] [19,20,21,22] Kondo & Sugimoto (2022) [16] Vol.29, No.3, 2024 Kondo & Sugimoto (2023) [23] [9, 16,17] [ 19,23] 0% 100% [15,16] [19] ...

Reference:

Full-body illusion of avatars with different transparency
Effects on Size Perception by Changing Dynamic Invisible Body Size
  • Citing Conference Paper
  • March 2024

... Studies have shown that sharing one's FOV with partners can significantly improve the efficiency of collaborative tasks by providing contextual and visual information [6,15,69]. Supporting multiple parallel views in a single user task also allows users to identify and localize target objects out of view more accurately in VR [61,62]. However, these studies do not explore the ubiquitous yet fundamental task of objects' delivery or transfer in the shared view, and most of the proposed techniques only focus on recognizing the partner's point of interest. ...

Evaluations of Parallel Views for Sequential VR Search Tasks
  • Citing Conference Paper
  • April 2024

... The questionnaire included items related to the sense of agency and immersion, with the sense of agency referring to the feeling of being able to control a seen object-in this case, the drone. This concept is explored in studies of body awareness in both actual [7,8] and VR [9][10][11] settings. ...

Investigating the perceptual attribution of a virtual robotic limb synchronizing with hand and foot simultaneously

Frontiers in Virtual Reality

... This model employs graph convolutions to characterize the interactions among on-road objects. Nitta et al. [30] develop a model that extracts temporal features from optical flow images to infer the states of moving objects. The optical flow images are also used in Malla et al. [26] to assess the states of moving objects. ...

Importance Rank-Learning of Objects in Urban Scenes for Assisting Visually Impaired People

IEEE Access

... The work supports visualized interactivity with remote participants. Niwa et al. state the importance of AI agents to augment human cognitive abilities [3]. Their approach is to implement a similar-looking AI avatar as a participant to investigate the trustworthiness level of agent statements. ...

Investigating Effects of Facial Self-Similarity Levels on the Impression of Virtual Agents in Serious/Non-Serious Contexts
  • Citing Conference Paper
  • March 2023

... We can extract at least three different strategies to control the additional DOFs, based on analogous reflections in the field of robotic [24]. First , a transfer control maps the additional DOFs to other real limbs DOFs [32,43,57], the remapping can lead one DOF of the real body to control several virtual DOFs (e.g., the movement of one single real hand is mapped to several virtual hands' movement [99]). The second strategy, augmentation control, consists in mapping additional DOFs to user's inputs other than their movements, such as their voice [116], or neural signals [89]. ...

Sensory Attenuation With a Virtual Robotic Arm Controlled Using Facial Movements
  • Citing Article
  • February 2023

IEEE Transactions on Visualization and Computer Graphics

... Tracking systems have a critical role in spatial computing, facilitating smooth user interactions with digital content within the physical environment [26]. These systems are necessary for precisely positioning virtual objects, interpreting user actions, and ensuring an immersive experience. ...

Tracking Systems: Calibration, Hardware, and Peripherals
  • Citing Chapter
  • January 2023

... The pre-treatment and normalization of landmarks successfully reduced data variance, leading to improved performance. Importantly, all the algorithms were trained and tested in real-time, highlighting their potential for practical applications beyond the academic environment, as also indicated in [4,29]. The promising results of our study suggest that simple normalization techniques can enhance the analysis of human emotions using images, as supported by [1,5]. ...

Consistent Smile Intensity Estimation from Wearable Optical Sensors
  • Citing Conference Paper
  • October 2022

... • Articulation Extension: a dissimilarity where a joint rotation limitation is extended, either by increasing its DOF or maximum rotation amplitudes. Such dissimilarity allows impossible postures, such as directing arms backward [39], or spreading fingers in an unrealistic manner [82]. ...

Embodiment of an Avatar with Unnatural Arm Movements
  • Citing Conference Paper
  • October 2022