DIGITAL IMAGE ANALYSIS USING MULTI-STEP ANALYSIS
First Claim
1. A computer-implemented feature extraction method for classifying pixels of a digitized pathology image, the method to be performed by a system comprising at least one processor and at least one memory, the method comprising:
- generating a plurality of characterized pixels from a digitized pathology image;
determining by the system in a first step feature analysis a first region and a first remainder region of the digitized pathology image based on the plurality of characterized pixels;
determining by the system, in a plurality of subsequent feature analysis steps, subsequent regions and subsequent remainder regions, wherein each feature analysis step determines a corresponding image region and a corresponding remainder region based on a remainder region determined by an earlier feature analysis step; and
classifying by the system part or all of the digitized pathology image based on the determined first region, and the determined subsequent regions.
1 Assignment
0 Petitions
Accused Products
Abstract
Systems and methods for implementing a multi-step image recognition framework for classifying digital images are provided. The provided multi-step image recognition framework utilizes a gradual approach to model training and image classification tasks requiring multi-dimensional ground truths. A first step of the multi-step image recognition framework differentiates a first image region from a remainder image region. Each subsequent step operates on a remainder image region from the previous step. The provided multi-step image recognition framework permits model training and image classification tasks to be performed more accurately and in a less resource intensive fashion than conventional single-step image recognition frameworks.
24 Citations
18 Claims
-
1. A computer-implemented feature extraction method for classifying pixels of a digitized pathology image, the method to be performed by a system comprising at least one processor and at least one memory, the method comprising:
-
generating a plurality of characterized pixels from a digitized pathology image; determining by the system in a first step feature analysis a first region and a first remainder region of the digitized pathology image based on the plurality of characterized pixels; determining by the system, in a plurality of subsequent feature analysis steps, subsequent regions and subsequent remainder regions, wherein each feature analysis step determines a corresponding image region and a corresponding remainder region based on a remainder region determined by an earlier feature analysis step; and classifying by the system part or all of the digitized pathology image based on the determined first region, and the determined subsequent regions. - View Dependent Claims (2, 3, 6, 7, 8, 9)
-
-
4. The method of claim wherein,
the digital pathology image is a training image; -
each pixel of the plurality of pixels is labeled with a ground truth; each first layer model from among the plurality of first layer models is generated by machine-learning algorithms based on a correspondence between the ground truth of each pixel of the plurality of pixels and the feature descriptor values of the feature of each pixel of the plurality of pixels corresponding to a designated feature type from among the plurality of feature types; and a different first layer model is generated to correspond to each feature type from among the plurality of feature types. - View Dependent Claims (5)
-
-
10. A system or image recognition analysis of a digital image comprising:
-
a memory having program instructions and data storage space; a processor configured to use the program instructions to perform the steps of; generating a plurality of characterized pixels from a digitized pathology image; determining in a first step feature analysis a first region and a first remainder region of the digitized pathology image based on the plurality of characterized pixels; determining in a plurality of subsequent feature analysis steps subsequent regions and subsequent remainder regions, wherein each feature analysis steps determines a corresponding image region and a corresponding remainder region based on a remainder region determined by an earlier feature analysis step; and classifying part or all of the digitized pathology image based on the determined first region, and the determined subsequent regions. - View Dependent Claims (11, 12, 13, 14, 15, 16, 17, 18)
-
Specification