Migration & Analysis of Images
Business Problem
- Our client had ECG data stored in fragmented data storages
- The goal was to create a data lake, to enable image processing capabilities from an unified data pipeline
- The efficiency provided by the legacy system was not up to the mark
Abzooba’s Solution:
- Flexible On-Prem Hadoop environment for data lake
- Integration of multiple sources for data ingestion into data lake, in multiple formats, supporting ad hoc analysis
- Spark-Scala was used to process the DICOM images and the team also developed and published a 3rd party Spark library to aid this task
- Search interface for users to extract DICOM images and perform analysis.
- Generation of relevant metadata for each ECG image using machine learning and image processing components
Business Benefits:
- Real-time access to:
- patient demographic data
- textual information
- measurement values
- DICOM image data
- Increase efficiency of processing huge volumes of image metadata by 70%
Tech Stack
