CAMEL: A Weakly Supervised Learning Framework for Histopathology Image Segmentation

Since the end of 2019, the COVID-19, which has swept across the world, has caused serious impacts on public health and economy. Although Reverse Transcription-Polymerase Chain Reaction (RT-PCR) is the gold standard for clinical diagnosis, it is very time-consuming and labor-intensive. At the same time, more and more people have doubted the sensitivity of RT-PCR. Therefore, Computed Tomography (CT) images are used as a substitute for RT-PCR. Powered by the research of the field of artificial intelligence, deep learning, which is a branch of machine learning, has made a great success on medical image segmentation. However, general full supervision methods require pixel-level point-by-point annotations, which is very costly. In this paper, we put forward an image segmentation method based on weakly supervised learning for CT images of COVID-19, which can effectively segment the lung infection area and doesn’t require pixel-level labels. Our method is contrasted with another four weakly supervised learning methods in recent years, and the results have been significantly improved.

Download Full-text

A Weakly Supervised Learning Framework for Detecting Social Anxiety and Depression

Proceedings of the ACM on Interactive Mobile Wearable and Ubiquitous Technologies ◽

10.1145/3214284 ◽

2018 ◽

Vol 2 (2) ◽

pp. 1-26 ◽

Cited By ~ 8

Author(s):

Asif Salekin ◽

Jeremy W. Eberle ◽

Jeffrey J. Glenn ◽

Bethany A. Teachman ◽

John A. Stankovic

Keyword(s):

Social Anxiety ◽

Supervised Learning ◽

Anxiety And Depression ◽

Weakly Supervised Learning ◽

Learning Framework ◽

Weakly Supervised

Download Full-text

Figure-Ground Image Segmentation Helps Weakly-Supervised Learning of Objects

Computer Vision – ECCV 2010 - Lecture Notes in Computer Science ◽

10.1007/978-3-642-15567-3_41 ◽

2010 ◽

pp. 561-574 ◽

Cited By ~ 3

Author(s):

Katerina Fragkiadaki ◽

Jianbo Shi

Keyword(s):

Image Segmentation ◽

Supervised Learning ◽

Weakly Supervised Learning ◽

Weakly Supervised

Download Full-text

Tag, Copy or Predict: A Unified Weakly-Supervised Learning Framework for Visual Information Extraction using Sequences

Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence ◽

10.24963/ijcai.2021/150 ◽

2021 ◽

Author(s):

Jiapeng Wang ◽

Tianwei Wang ◽

Guozhi Tang ◽

Lianwen Jin ◽

Weihong Ma ◽

...

Keyword(s):

Information Extraction ◽

Supervised Learning ◽

Visual Information ◽

Character Recognition ◽

Optical Character Recognition ◽

Input Sequence ◽

Time Step ◽

Weakly Supervised Learning ◽

Learning Framework ◽

Weakly Supervised

Visual information extraction (VIE) has attracted increasing attention in recent years. The existing methods usually first organized optical character recognition (OCR) results in plain texts and then utilized token-level category annotations as supervision to train a sequence tagging model. However, it expends great annotation costs and may be exposed to label confusion, the OCR errors will also significantly affect the final performance. In this paper, we propose a unified weakly-supervised learning framework called TCPNet (Tag, Copy or Predict Network), which introduces 1) an efficient encoder to simultaneously model the semantic and layout information in 2D OCR results, 2) a weakly-supervised training method that utilizes only sequence-level supervision; and 3) a flexible and switchable decoder which contains two inference modes: one (Copy or Predict Mode) is to output key information sequences of different categories by copying a token from the input or predicting one in each time step, and the other (Tag Mode) is to directly tag the input sequence in a single forward pass. Our method shows new state-of-the-art performance on several public benchmarks, which fully proves its effectiveness.

Download Full-text