Chemical Substance Response Prediction From Tissue Image Tiles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional patient response prediction techniques for pharmaceutical drugs are inaccurate and computationally inefficient due to the large size of biological images, requiring billions of parameters and failing to extract sufficient features from medical images, leading to prediction accuracy ranging from 20%-45%.
Innovation Solution
Implementing image processing and machine learning techniques to process biological images in a computationally efficient manner by generating image tiles, preprocessing them to identify discrete tissue components, and using artificial neural networks to predict patient responses based on learned predictive power of these components.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional patient response prediction techniques are used, then computational resources are consumed with billions of parameters, but prediction accuracy remains low (20%-45%) and computational efficiency is poor
Solution Approach 1:
The patent divides the large biological image into multiple smaller image tiles (e.g., 224x224 pixels each). This segmentation allows the system to process manageable portions of the image through the neural network, reducing computational complexity while maintaining the ability to extract comprehensive features from the entire tissue sample. The segmentation principle directly addresses the contradiction by enabling accurate predictions without requiring billions of parameters.
Solution Approach 2:
The patent extracts and processes only the relevant visual features from the biological images using image processing techniques and machine learning models. By extracting meaningful representations from the image tiles and focusing computational resources on these extracted features rather than processing the entire image with billions of parameters, the system achieves high prediction accuracy with reduced computational requirements.
2Productivity
If traditional techniques process large biological images, then computational efficiency is poor, but the system fails to extract sufficient features for accurate prediction
Solution Approach 1:
By segmenting the large biological image into smaller tiles, the system improves computational efficiency by processing smaller data chunks that can be analyzed more quickly and with fewer resources. Each tile maintains sufficient information content, preventing loss of critical features while enabling efficient processing.
Solution Approach 2:
The patent transforms the two-dimensional large biological image into multiple smaller two-dimensional tiles, and then processes these tiles through machine learning models that operate in a transformed feature space. This dimensional transformation allows efficient extraction of meaningful features without losing information, as the tiles collectively capture all spatial relationships and patterns present in the original image.
3Measurement precision
If the system processes biological images to predict patient responses, then prediction accuracy increases, but computational requirements and processing time increase
Solution Approach 1:
Segmenting the biological image into smaller tiles enables parallel processing and reduces the computational time required for each prediction. The system can process multiple tiles simultaneously or in sequence, significantly reducing total processing time compared to analyzing the entire large image at once, while maintaining high prediction accuracy through comprehensive feature extraction from all tiles.
Solution Approach 2:
The patent incorporates preliminary image processing steps that prepare and preprocess the image tiles before they are input to the machine learning model. This preliminary action includes techniques such as normalization, augmentation, and feature extraction that can be performed efficiently in advance, reducing the computational burden during the actual prediction phase and overall decreasing processing time while maintaining accuracy.
Data Source
AI summary
In an aspect, a data processing system includes a computer-readable memory comprising computer-executable instructions, and at least one processor configured to execute executable logic including at least one artificial neural network trained to predict one or more responses to a chemical substance by identifying one or more discrete biological tissue components in a biological image. When the at least one processor is executing the computer-executable instructions, the at least one processor is configured to carry out operations including: receiving spatially arranged image data representing a biological image of a patient; generating spatially arranged image tile data representing a plurality of image tiles; processing the spatially arranged image tile data through one or more data structures storing one or more portions of executable logic included in the artificial neural network to predict one or more responses of a patient.