Parsing human skeletons in an operating room

Belagiannis, Vasileios; Wang, Xinchao; Shitrit, Horesh Beny; Hashimoto, Kiyoshi; Stauder, Ralf; Aoki, Yoshimitsu; Kranzfelder, Michael; Schneider, Armin; Fua, Pascal; Ilić, Slobodan; Feußner, H.; Navab, Nassir

doi:10.1007/s00138-016-0792-4

Cited by 35 publications

(18 citation statements)

References 53 publications

(78 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The automatic recognition of activities during real surgeries to develop intelligent context-aware assistance systems is a recent field that has started to gain traction in the medical as well as computer vision community [15,16,17]. Work on analyzing humans in OR videos have generally focused on person bounding box detection and on human pose estimation, using either RGB or RGB-D data [18,19,20]. So far, face detection in the OR has however received very little attention, besides the recent work [21], described below.…”

Section: Introductionmentioning

confidence: 99%

Face detection in the operating room: comparison of state-of-the-art methods and a self-supervised approach

et al. 2019

View full text Add to dashboard Cite

Purpose: Face detection is a needed component for the automatic analysis and assistance of human activities during surgical procedures. Efficient face detection algorithms can indeed help to detect and identify the persons present in the room, and also be used to automatically anonymize the data. However, current algorithms trained on natural images do not generalize well to the operating room (OR) images. In this work, we provide a comparison of state-of-the-art face detectors on OR data and also present an approach to train a face detector for the OR by exploiting non-annotated OR images. Methods: We propose a comparison of 6 state-of-the-art face detectors on clinical data using Multi-View Operating Room Faces (MVOR-Faces), a dataset of operating room images capturing real surgical activities. We then propose to use self-supervision, a domain adaptation method, for the task of face detection in the OR. The approach makes use of non-annotated images to fine-tune a state-of-the-art detector for the OR without using any human supervision. Results: The results show that the best model, namely the tiny face detector, yields an average precision of 0.556 at Intersection over Union (IoU) of 0.5. Our self-supervised model using non-annotated clinical data outperforms this result by 9.2%. Conclusion: We present the first comparison of state-of-the-art face detectors on operating room images and show that results can be significantly improved by using self-supervision on non-annotated data.

show abstract

Section: Introductionmentioning

confidence: 99%

Face detection in the operating room: comparison of state-of-the-art methods and a self-supervised approach

et al. 2019

View full text Add to dashboard Cite

show abstract

“…These systems, which have a huge potential to improve safety and care, all rely on machine intelligence using computer vision models to extract semantic information from visual data. In particular, human detection and pose estimation in the operating room [1,6,8] is one of the key components to develop such [17] down-sampled at different resolutions using bicubic interpolation (resized for better visualization). Low-resolution depth images contain little information for the identification of patient and health professionals.…”

Section: Introductionmentioning

confidence: 99%

Human Pose Estimation on Privacy-Preserving Low-Resolution Depth Images

Srivastav

Gangi

Padoy

2019

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Human pose estimation (HPE) is a key building block for developing AI-based context-aware systems inside the operating room (OR). The 24/7 use of images coming from cameras mounted on the OR ceiling can however raise concerns for privacy, even in the case of depth images captured by RGB-D sensors. Being able to solely use lowresolution privacy-preserving images would address these concerns and help scale up the computer-assisted approaches that rely on such data to a larger number of ORs. In this paper, we introduce the problem of HPE on low-resolution depth images and propose an end-to-end solution that integrates a multi-scale super-resolution network with a 2D human pose estimation network. By exploiting intermediate feature-maps generated at different super-resolution, our approach achieves body pose results on low-resolution images (of size 64x48) that are on par with those of an approach trained and tested on full resolution images (of size 640x480).

show abstract

“…Multi-Object tracking (MOT) offers extensive applications ranges such as behavior and pose analysis, medical image analysis, and video surveillance [1]. MOT provides other difficulties like data association when comparing with the single-object tracking that aims to relate the detections over the frames to safe the identities of the targets.…”

Section: Introductionmentioning

confidence: 99%

An Efficient and Robust Multi-Object Recognition and Tracking Algorithm using Mask Region based Convolution Neural Network (R-CNN)

Nirmala¹,

Arivalagan²,

Arunkumar³

2019

IJITEE

View full text Add to dashboard Cite

Presently, Multi-Object tracking (MOT) is mainly applied for predicting the positions of many predefined objects across many successive frames with the provided ground truth position of the target in the first frame. The area of MOT gains more interest in the area of computer vision because of its applicability in various fields. Many works have been presented in recent years that intended to design a MOT algorithm with maximum accuracy and robustness. In this paper, we introduce an efficient as well as robust MOT algorithm using Mask R-CNN. The usage of Mask R-CNN effectively identifies the objects present in the image while concurrently creating a high-quality segmentation mask for every instance. The presented MOT algorithm is validated using three benchmark dataset and the results are extensive simulation. The presented tracking algorithm shows its efficiency to track multiple objects precisely

show abstract

Parsing human skeletons in an operating room

Cited by 35 publications

References 53 publications

Face detection in the operating room: comparison of state-of-the-art methods and a self-supervised approach

Face detection in the operating room: comparison of state-of-the-art methods and a self-supervised approach

Human Pose Estimation on Privacy-Preserving Low-Resolution Depth Images

An Efficient and Robust Multi-Object Recognition and Tracking Algorithm using Mask Region based Convolution Neural Network (R-CNN)

Contact Info

Product

Resources

About