RECORD
SPDA: Superpixel-based Data Augmentation for Biomedical Image Segmentation
Computer Vision and Pattern Recognition · Submitted 28 Feb 2019
abstract compiled · 2026-09-23 · no language-model review
The authors' main claim Our experiments of SPDA on four biomedical image datasets show that SPDA is effective and can consistently improve the performance of state-of-the-art fully convolutional networks for biomedical image segmentation in 2D and 3D images.
The compiled record
The paper's own sentences, arranged. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.
Open the compiled record
The abstract, sentence by sentence
Supervised training a deep neural network aims to "teach" the network to mimic human visual perception that is represented by image-and-label pairs in the training data. Superpixelized (SP) images are visually perceivable to humans, but a conventionally trained deep learning model often performs poorly when working on SP images. To better mimic human visual perception, we think it is desirable for the deep learning model to be able to perceive not only raw images but also SP images. In this paper, we propose a new superpixel-based data augmentation (SPDA) method for training deep learning models for biomedical image segmentation. Our method applies a superpixel generation scheme to all the original training images to generate superpixelized images. The SP images thus obtained are then jointly used with the original training images to train a deep learning model. Our experiments of SPDA on four biomedical image datasets show that SPDA is effective and can consistently improve the performance of state-of-the-art fully convolutional networks for biomedical image segmentation in 2D and 3D images. Additional studies also demonstrate that SPDA can practically reduce the generalization gap.
background contribution method result interpretation main claim
| # | sentence | role | the authors' own work | states the main conclusion | strength |
|---|---|---|---|---|---|
| s0 | Supervised training a deep neural network aims to "teach" the network to mimic human visual perception that is represented by image-and-label pairs in the training data. | background 1.00 | 0.05 | 0.02 | asserted 1.00 |
| s1 | Superpixelized (SP) images are visually perceivable to humans, but a conventionally trained deep learning model often performs poorly when working on SP images. | background 0.72 | 0.26 | 0.02 | asserted 0.86 |
| s2 | To better mimic human visual perception, we think it is desirable for the deep learning model to be able to perceive not only raw images but also SP images. | background 0.57 | 0.09 | 0.04 | speculative 0.57 |
| s3 | In this paper, we propose a new superpixel-based data augmentation (SPDA) method for training deep learning models for biomedical image segmentation. | contribution 0.99 | 0.85 | 0.08 | asserted 0.99 |
| s4 | Our method applies a superpixel generation scheme to all the original training images to generate superpixelized images. | method 0.99 | 0.93 | 0.02 | asserted 1.00 |
| s5 | The SP images thus obtained are then jointly used with the original training images to train a deep learning model. | method 1.00 | 0.89 | 0.04 | asserted 0.99 |
| s6 | Our experiments of SPDA on four biomedical image datasets show that SPDA is effective and can consistently improve the performance of state-of-the-art fully convolutional networks for biomedical image segmentation in 2D and 3D images. | result 1.00 | 0.97 | 0.83 | asserted 1.00 |
| s7 | Additional studies also demonstrate that SPDA can practically reduce the generalization gap. | result 0.89 | 0.90 | 0.14 | asserted 0.99 |
What was examined, and what was not
- arXiv listing abstract v1
- full text
- figures and tables
- references and citing sentences
- relations to other papers
- P abstract
- S sections
- T body sentences
- C citations
- R cross-paper relations
The paper
Original abstract
Supervised training a deep neural network aims to "teach" the network to mimic human visual perception that is represented by image-and-label pairs in the training data. Superpixelized (SP) images are visually perceivable to humans, but a conventionally trained deep learning model often performs poorly when working on SP images. To better mimic human visual perception, we think it is desirable for the deep learning model to be able to perceive not only raw images but also SP images. In this paper, we propose a new superpixel-based data augmentation (SPDA) method for training deep learning models for biomedical image segmentation. Our method applies a superpixel generation scheme to all the original training images to generate superpixelized images. The SP images thus obtained are then jointly used with the original training images to train a deep learning model. Our experiments of SPDA on four biomedical image datasets show that SPDA is effective and can consistently improve the performance of state-of-the-art fully convolutional networks for biomedical image segmentation in 2D and 3D images. Additional studies also demonstrate that SPDA can practically reduce the generalization gap.
Human review
How the literature uses it
What this paper does with the arXiv papers it cites
What it cites · 3 sentences without a confident label
Uses the method or tool of 2 papers
-
In Appendix [ref], using variational auto-encoders (VAEs) [ref] and the Kullback-Leibler divergence [ref] , we show that the training set with SPDA-generated data is closer to the test set in terms of the overall data distribution.
Auto-Encoding Variational Bayes · Technical considerations · 0.91
-
Adam [ref] optimizer is used for model optimization.
Adam: A Method for Stochastic Optimization · Experiments · 1.00
Compares with 1 paper
-
In particular, using SPDA, DCN can be trained to perform considerably better than the state-of-the-art model [ref] .
MILD-Net: Minimal Information Loss Dilated Network for Gland Instance Segmentation in Colon Histology Images · Experiments · 1.00
Cites as background 3 papers
-
Traditional data augmentation methods use a combination of geometric transformations to artificially inflate training data [ref] .
The Effectiveness of Data Augmentation in Image Classification using Deep Learning · Introduction · 0.99
-
As an algorithm based (non-learning based) data augmentation technique, mixup [ref] was proposed to generate new image samples “between” pairs of training samples for image classification problems.
mixup: Beyond Empirical Risk Minimization · Introduction · 0.95
-
Recently, generative adversarial networks (GANs) have been used for data augmentation (e.g., [ref] ).
Data Augmentation Generative Adversarial Networks · Introduction · 1.00
Discussion
Know something about this paper? Leave a comment, or highlight any passage above to comment on it.