An official website of the United States government
Annotating medical images for the purposes of training computer vision models is an extremely laborious task that takes time and resources away from expert clinicians. Active learning (AL) is a machine learning paradigm that mitigates this problem by deliberately proposing data points that should be labeled in order to maximize model performance. We propose a novel AL algorithm for segmentation, ALGES, that utilizes gradient embeddings to effectively select laparoscopic images to be labeled by some external oracle while reducing annotation effort. Given any unlabeled image, our algorithm treats predicted segmentations as truth and computes gradients with respect to the model parameters of the last layer in a segmentation network. The norms of these per-pixel gradient vectors correspond to the magnitude of the induced change in model parameters and contain rich information about the model’s predictive uncertainty. Our algorithm then computes gradients embeddings in two ways, and we employ a center-finding algorithm with these embeddings to procure representative and diverse batches in each round of AL. An advantage of our approach is extensibility to any model architecture and differentiable loss scheme for semantic segmentation. We apply our approach to a public data set of laparoscopic cholecystectomy images and show that it outperforms current AL algorithms in selecting the most informative data points for improving the segmentation model. Our code is available at https://github.com/josaklil-ai/surg-active-learning.
more »
« less
Active Learning Technique for Multimodal Brain Tumor Segmentation using Limited Labeled Images
Image segmentation is an essential step in biomedical image analysis. In recent years, deep learning models have achieved signi cant success in segmentation. However, deep learning requires the availability of large annotated data to train these models, which can be challenging in biomedical imaging domain. In this paper, we aim to accomplish biomedical image segmentation with limited labeled data using active learning. We present a deep active learning framework that selects additional data points to be annotated by combining U-Net with an efficient and effective query strategy to capture the most uncertain and representative points. This algorithm decouples the representative part by first finding the core points in the unlabeled pool and then selecting the most uncertain points from the reduced pool, which are different from the labeled pool. In our experiment, only 13% of the dataset was required with active learning to outperform the model trained on the entire 2018 MIC- CAI Brain Tumor Segmentation (BraTS) dataset. Thus, active learning reduced the amount of labeled data required for image segmentation without a signi cant loss in the accuracy.
more »
« less
- Award ID(s):
- 1838730
- PAR ID:
- 10119080
- Date Published:
- Journal Name:
- MICCAI Workshop on Medical Image Learning with Less Labels and Imperfect Data
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Accurate instrument segmentation in the endoscopic vision of minimally invasive surgery is challenging due to complex instruments and environments. Deep learning techniques have shown competitive performance in recent years. However, deep learning usually requires a large amount of labeled data to achieve accurate prediction, which poses a significant workload. To alleviate this workload, we propose an active learning-based framework to generate synthetic images for efficient neural network training. In each active learning iteration, a small number of informative unlabeled images are first queried by active learning and manually labeled. Next, synthetic images are generated based on these selected images. The instruments and backgrounds are cropped out and randomly combined with blending and fusion near the boundary. The proposed method leverages the advantage of both active learning and synthetic images. The effectiveness of the proposed method is validated on two sinus surgery datasets and one intraabdominal surgery dataset. The results indicate a considerable performance improvement, especially when the size of the annotated dataset is small. All the code is open-sourced at: https://github.com/HaonanPeng/active_syn_generatormore » « less
-
ABSTRACT Cochlear hair cell stereocilia bundles are key organelles required for normal hearing. Often, deafness mutations cause aberrant stereocilia heights or morphology that are visually apparent but challenging to quantify. Actin-based structures, stereocilia are easily and most often labeled with phalloidin then imaged with 3D confocal microscopy. Unfortunately, phalloidin non-specifically labels all the actin in the tissue and cells and therefore results in a challenging segmentation task wherein the stereocilia phalloidin signal must be separated from the rest of the tissue. This can require many hours of manual human effort for each 3D confocal image stack. Currently, there are no existing software pipelines that provide an end-to-end automated solution for 3D stereocilia bundle instance segmentation. Here we introduce VASCilia, a Napari plugin designed to automatically generate 3D instance segmentation and analysis of 3D confocal images of cochlear hair cell stereocilia bundles stained with phalloidin. This plugin combines user-friendly manual controls with advanced deep learning-based features to streamline analyses. With VASCilia, users can begin their analysis by loading image stacks. The software automatically preprocesses these samples and displays them in Napari. At this stage, users can select their desired range of z-slices, adjust their orientation, and initiate 3D instance segmentation. After segmentation, users can remove any undesired regions and obtain measurements including volume, centroids, and surface area. VASCilia introduces unique features that measures bundle heights, determines their orientation with respect to planar polarity axis, and quantifies the fluorescence intensity within each bundle. The plugin is also equipped with trained deep learning models that differentiate between inner hair cells and outer hair cells and predicts their tonotopic position within the cochlea spiral. Additionally, the plugin includes a training section that allows other laboratories to fine-tune our model with their own data, provides responsive mechanisms for manual corrections through event-handlers that check user actions, and allows users to share their analyses by uploading a pickle file containing all intermediate results. We believe this software will become a valuable resource for the cochlea research community, which has traditionally lacked specialized deep learning-based tools for obtaining high-throughput image quantitation. Furthermore, we plan to release our code along with a manually annotated dataset that includes approximately 55 3D stacks featuring instance segmentation. This dataset comprises a total of 1,870 instances of hair cells, distributed between 410 inner hair cells and 1,460 outer hair cells, all annotated in 3D. As the first open-source dataset of its kind, we aim to establish a foundational resource for constructing a comprehensive atlas of cochlea hair cell images. Together, this open-source tool will greatly accelerate the analysis of stereocilia bundles and demonstrates the power of deep learning-based algorithms for challenging segmentation tasks in biological imaging research. Ultimately, this initiative will support the development of foundational models adaptable to various species, markers, and imaging scales to advance and accelerate research within the cochlea research community.more » « less
-
Traumatic brain injury (TBI) is a massive public health problem worldwide. Accurate and fast automatic brain hematoma segmentation is important for TBI diagnosis, treatment and outcome prediction. In this study, we developed a fully automated system to detect and segment hematoma regions in head Computed Tomography (CT) images of patients with acute TBI. We first over-segmented brain images into superpixels and then extracted statistical and textural features to capture characteristics of superpixels. To overcome the shortage of annotated data, an uncertainty-based active learning strategy was designed to adaptively and iteratively select the most informative unlabeled data to be annotated for training a Support Vector Machine classifier (SVM). Finally, the coarse segmentation from the SVM classifier was incorporated into an active contour model to improve the accuracy of the segmentation. From our experiments, the proposed active learning strategy can achieve a comparable result with 5 times fewer labeled data compared with regular machine learning. Our proposed automatic hematoma segmentation system achieved an average Dice coefficient of 0.60 on our dataset, where patients are from multiple health centers and at multiple levels of injury. Our results show that the proposed method can effectively overcome the challenge of limited and highly varied dataset.more » « less
-
The COVID-19 pandemic has intensified the need for home-based cardiac health monitoring systems. Despite advancements in electrocardiograph (ECG) and phonocardiogram (PCG) wearable sensors, accurate heart sound segmentation algorithms remain understudied. Existing deep learning models, such as convolutional neural networks (CNN) and recurrent neural networks (RNN), struggle to segment noisy signals using only PCG data. We propose a two-step heart sound segmentation algorithm that analyzes synchronized ECG and PCG signals. The first step involves heartbeat detection using a CNN-LSTM-based model on ECG data, and the second step focuses on beat-wise heart sound segmentation with a 1D U-Net that incorporates multi-modal inputs. Our method leverages temporal correlation between ECG and PCG signals to enhance segmentation performance. To tackle the label-hungry issue in AI-supported biomedical studies, we introduce a segment-wise contrastive learning technique for signal segmentation, overcoming the limitations of traditional contrastive learning methods designed for classification tasks. We evaluated our two-step algorithm using the PhysioNet 2016 dataset and a private dataset from Bayland Scientific, obtaining a 96.43 F1 score on the former. Notably, our segment-wise contrastive learning technique demonstrated effective performance with limited labeled data. When trained on just 1% of labeled PhysioNet data, the model pre-trained on the full unlabeled dataset only dropped 2.88 in the F1 score, outperforming the SimCLR method. Overall, our proposed algorithm and learning technique present promise for improving heart sound segmentation and reducing the need for labeled data.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- The DOI is not currently available.
