Smartphone based automatic organ validation in ultrasound video

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Telemedicine

Background:

• Telesonography faces challenges due to a lack of trained sonographers in remote areas, leading to inadequate ultrasound video quality for diagnosis.
• Existing video transmission methods are insufficient for transmitting valid medical videos, necessitating mHealth solutions.
• Smartphone technology advancements enable sophisticated medical image processing on mobile devices.

Purpose of the Study:

• To develop an organ validation algorithm for evaluating ultrasound video content.
• To create a smartphone application that automatically detects valid ultrasound frames and discards invalid ones.
• To guide semi-skilled personnel in acquiring representative patient data for remote diagnosis.

Main Methods:

• Extraction of GIST features from the Region of Interest (ROI) of ultrasound frames.
• Classification of frames as valid or invalid using a Support Vector Machine (SVM) classifier with a quadratic kernel.
• Development of a smartphone application to automate frame validation and video compression.

Main Results:

• The developed smartphone application achieved an accuracy of 94.93% in classifying valid and invalid ultrasound frames.
• The system effectively identifies frames with clear organ visibility and ignores those without.
• The application produces a compressed-sized video of validated frames.

Conclusions:

• The proposed organ validation algorithm and smartphone application can significantly improve the quality of ultrasound videos transmitted from remote areas.
• This mHealth solution empowers semi-skilled individuals to acquire diagnostically relevant ultrasound data.
• The developed system offers a practical approach to enhance remote medical diagnosis through automated video quality assessment.