Parallel multiple instance learning for extremely large histopathology image analysis

Abstract

Background: Histopathology images are critical for medical diagnosis, e.g., cancer and its treatment. A standard histopathology slice can be easily scanned at a high resolution of, say, 200,000×200,000 pixels. These high resolution images can make most existing imaging processing tools infeasible or less effective when operated on a single machine with limited memory, disk space and computing power.

Results: In this paper, we propose an algorithm tackling this new emerging "big data" problem utilizing parallel computing on High-Performance-Computing (HPC) clusters. Experimental results on a large-scale data set (1318 images at a scale of 10 billion pixels each) demonstrate the efficiency and effectiveness of the proposed algorithm for low-latency real-time applications.

Conclusions: The framework proposed an effective and efficient system for extremely large histopathology image analysis. It is based on the multiple instance learning formulation for weakly-supervised learning for image classification, segmentation and clustering. When a max-margin concept is adopted for different clusters, we obtain further improvement in clustering performance.

Keywords: Histopathology image analysis; Microscopic image analysis; Multiple instance learning; Parallelization.

Fig. 1

Parallel Multiple Instance Learning (P-MIL) on High-Performance-Computing (HPC) cluster. Red: positive instances; Green: negative instances. At first, we divide and distribute data to the nodes. The master will collect the results calculated by individual nodes, train multiple classifiers and choose the best one. Next, the slaves receive the best weak classifier and calculate an individual α value. The master node then will synchronize all the nodes, choose the α _best and broadcast it. At last, all the nodes will update classifiers with the α _best and update new clusters with the new classifiers through communication, in which the master will coordinate to ensure data coherence. The program will continue running in a loop until the loop ends

Fig. 2

Illustrations of max-margin using linear classifier. Green, red and purple dots represent three specific cancer subtypes, while black dots represent non-cancer instances. Linear boundaries are trained to separate cancer subtypes from each other (intra-class) and the non-cancer (inter-class)

Fig. 3

Illustrations of cluster competition using max-margin linear classifier. Green and red dots represent two classes. In a, two classes are initialized by K-means method. In b–d, cluster competition takes place until the model converges. Specifically, instances in each class are classified by linear classifiers, according to which they update their labels. Then, a new classifier is trained based on the new labels. The cluster competition converges when both classifiers and labels of instances become in a stable state

Fig. 4

The mirrored Receiver Operating Characteristic (ROC) curve for comparisons of piece-level classification results with Multiple Instance Learning (MIL), Multiple Clustered Instance Learning (MCIL) and Parallel Multiple Instance Learning (P-MIL). The generalized mean (GM) model is the soft-max function in the methods

Fig. 5

Image Types: a: The original images. b–d: The instance-level segmentations for MIL, MCIL and P-MIL respectively. e: The ground truth. ANC: abnormal; NC: normal. Different colors represent different cancer subtypes