Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

Multimodal deep learning for predicting WHO/ISUP grading in renal tumors on CT using a self-attention-based model: variable Vision Transformer (vViT).

European journal of radiology open·2026

Same author

Fatal Injuries Caused by Falling Trees: Correlation Between Postmortem Computed Tomography and Forensic Autopsy Findings in Eight Cases.

International journal of legal medicine·2026

Same author

Heat hematoma-like epidural collection mimicking traumatic hematoma in a reheated bathtub death.

Journal of forensic sciences·2026

Same author

Evaluation of Optimal Ordered Subset Expectation Maximization Reconstruction for [I-123]FP-CIT Single-Photon Emission Computed Tomography Using a Ring-Type Semiconductor Scanner.

The Tohoku journal of experimental medicine·2026

Same author

Transformer-based Deep Learning Models with Shape Guidance for Predicting Breast Cancer in Mammography Images.

Journal of imaging informatics in medicine·2025

Same author

Bilateral Information-Guided Diagnosis of Breast Masses in Mammography Using Vision Transformer.

IEEE journal of biomedical and health informatics·2025

Search research articles

Related Experiment Video

Updated: Aug 19, 2025

Image Rendering Techniques in Postmortem Computed Tomography: Evaluation of Biological Health and Profile in Stranded Cetaceans

Published on: September 27, 2020

A 2.5D Deep Learning-Based Method for Drowning Diagnosis Using Post-Mortem Computed Tomography.

Yuwen Zeng, Xiaoyong Zhang, Yusuke Kawasumi

IEEE Journal of Biomedical and Health Informatics

|November 29, 2022

Summary

Diagnosing drowning via autopsy is difficult. A novel 2.5D deep learning method enhances computer-aided diagnosis (CAD) using multi-slice computed tomography (MSCT) scans, improving accuracy and identifying key drowning indicators.

More Related Videos

A Volumetric Method for Quantification of Cerebral Vasospasm in a Murine Model of Subarachnoid Hemorrhage

A Volumetric Method for Quantification of Cerebral Vasospasm in a Murine Model of Subarachnoid Hemorrhage

Published on: July 28, 2018

Combined Near-infrared Fluorescent Imaging and Micro-computed Tomography for Directly Visualizing Cerebral Thromboemboli

Combined Near-infrared Fluorescent Imaging and Micro-computed Tomography for Directly Visualizing Cerebral Thromboemboli

Published on: September 25, 2016

Related Experiment Videos

Last Updated: Aug 19, 2025

Image Rendering Techniques in Postmortem Computed Tomography: Evaluation of Biological Health and Profile in Stranded Cetaceans

Published on: September 27, 2020

A Volumetric Method for Quantification of Cerebral Vasospasm in a Murine Model of Subarachnoid Hemorrhage

A Volumetric Method for Quantification of Cerebral Vasospasm in a Murine Model of Subarachnoid Hemorrhage

Published on: July 28, 2018

Combined Near-infrared Fluorescent Imaging and Micro-computed Tomography for Directly Visualizing Cerebral Thromboemboli

Combined Near-infrared Fluorescent Imaging and Micro-computed Tomography for Directly Visualizing Cerebral Thromboemboli

Published on: September 25, 2016

Area of Science:

Forensic pathology
Medical imaging
Artificial intelligence

Background:

Diagnosing drowning in autopsies is challenging due to complex pathophysiology and limited forensic radiology expertise.
Existing computer-aided diagnosis (CAD) systems often rely on 2D data, neglecting crucial 3D information from CT scans.
Traditional 3D deep learning methods demand substantial data and computational resources.

Purpose of the Study:

To develop an effective computer-aided diagnosis (CAD) system for drowning detection using post-mortem multi-slice computed tomography (MSCT) data.
To propose a novel 2.5D deep learning approach that balances 3D data representation with computational efficiency.
To enhance the explainability of AI-driven diagnostic results through visualization techniques.

Main Methods:

A 2.5D deep learning method was developed, converting 3D MSCT data into representative 2D images for training.
The 2.5D approach utilizes a data subset to capture essential case information while minimizing redundancy.
Gradient-weighted Class Activation Mapping (Grad-CAM) was employed for visualizing image features indicative of drowning.

Main Results:

The proposed 2.5D method demonstrated superior performance in drowning diagnosis compared to conventional 2D, previous 2.5D, and 3D deep learning models.
Evaluation was conducted on an MSCT dataset from Tohoku University.
Visual assessment confirmed the method's ability to identify relevant "saliency regions" associated with drowning in CT images.

Conclusions:

The developed 2.5D deep learning method offers a promising solution for improving the accuracy and efficiency of computer-aided drowning diagnosis from MSCT scans.
This approach effectively leverages 3D CT data while mitigating the computational burdens of traditional 3D methods.
Explainable AI, through visualization, aids in understanding the diagnostic basis, supporting forensic specialists.