Self-supervised out-of-distribution detection in wireless capsule endoscopy images.

While deep learning has displayed excellent performance in a broad spectrum of application areas, neural networks still struggle to recognize what they have not seen, i.e., out-of-distribution (OOD) inputs. In the medical field, building robust models that are able to detect OOD images is highly critical, as these rare images could show diseases or anomalies that should be detected. In this study, we use wireless capsule endoscopy (WCE) images to present a novel patch-based self-supervised approach comprising three stages. First, we train a triplet network to learn vector representations of WCE image patches. Second, we cluster the patch embeddings to group patches in terms of visual similarity. Third, we use the cluster assignments as pseudolabels to train a patch classifier and use the Out-of-Distribution Detector for Neural Networks (ODIN) for OOD detection. The system has been tested on the Kvasir-capsule, a publicly released WCE dataset. Empirical results show an OOD detection improvement compared to baseline methods. Our method can detect unseen pathologies and anomalies such as lymphangiectasia, foreign bodies and blood with AUROC>0.6. This work presents an effective solution for OOD detection models without needing labeled images.

Anatomical landmarks localization for capsule endoscopy studies.

Wireless Capsule Endoscopy is a medical procedure that uses a small, wireless camera to capture images of the inside of the digestive tract. The identification of the entrance and exit of the small bowel and of the large intestine is one of the first tasks that need to be accomplished to read a video. This paper addresses the design of a clinical decision support tool to detect these anatomical landmarks. We have developed a system based on deep learning that combines images, timestamps, and motion data to achieve state-of-the-art results. Our method does not only classify the images as being inside or outside the studied organs, but it is also able to identify the entrance and exit frames. The experiments performed with three different datasets (one public and two private) show that our system is able to approximate the landmarks while achieving high accuracy on the classification problem (inside/outside of the organ). When comparing the entrance and exit of the studied organs, the distance between predicted and real landmarks is reduced from 1.5 to 10 times with respect to previous state-of-the-art methods.

Time-based self-supervised learning for Wireless Capsule Endoscopy.

State-of-the-art machine learning models, and especially deep learning ones, are significantly data-hungry; they require vast amounts of manually labeled samples to function correctly. However, in most medical imaging fields, obtaining said data can be challenging. Not only the volume of data is a problem, but also the imbalances within its classes; it is common to have many more images of healthy patients than of those with pathology. Computer-aided diagnostic systems suffer from these issues, usually over-designing their models to perform accurately. This work proposes using self-supervised learning for wireless endoscopy videos by introducing a custom-tailored method that does not initially need labels or appropriate balance. We prove that using the inferred inherent structure learned by our method, extracted from the temporal axis, improves the detection rate on several domain-specific applications even under severe imbalance. State-of-the-art results are achieved in polyp detection, with 95.00 ± 2.09% Area Under the Curve, and 92.77 ± 1.20% accuracy in the CAD-CAP dataset.

Artificial intelligence to improve polyp detection and screening time in colon capsule endoscopy.

Colon Capsule Endoscopy (CCE) is a minimally invasive procedure which is increasingly being used as an alternative to conventional colonoscopy. Videos recorded by the capsule cameras are long and require one or more experts' time to review and identify polyps or other potential intestinal problems that can lead to major health issues. We developed and tested a multi-platform web application, AI-Tool, which embeds a Convolution Neural Network (CNN) to help CCE reviewers. With the help of artificial intelligence, AI-Tool is able to detect images with high probability of containing a polyp and prioritize them during the reviewing process. With the collaboration of 3 experts that reviewed 18 videos, we compared the classical linear review method using RAPID Reader Software v9.0 and the new software we present. Applying the new strategy, reviewing time was reduced by a factor of 6 and polyp detection sensitivity was increased from 81.08 to 87.80%.

WCE polyp detection with triplet based embeddings.

Wireless capsule endoscopy is a medical procedure used to visualize the entire gastrointestinal tract and to diagnose intestinal conditions, such as polyps or bleeding. Current analyses are performed by manually inspecting nearly each one of the frames of the video, a tedious and error-prone task. Automatic image analysis methods can be used to reduce the time needed for physicians to evaluate a capsule endoscopy video. However these methods are still in a research phase. In this paper we focus on computer-aided polyp detection in capsule endoscopy images. This is a challenging problem because of the diversity of polyp appearance, the imbalanced dataset structure and the scarcity of data. We have developed a new polyp computer-aided decision system that combines a deep convolutional neural network and metric learning. The key point of the method is the use of the Triplet Loss function with the aim of improving feature extraction from the images when having small dataset. The Triplet Loss function allows to train robust detectors by forcing images from the same category to be represented by similar embedding vectors while ensuring that images from different categories are represented by dissimilar vectors. Empirical results show a meaningful increase of AUC values compared to state-of-the-art methods. A good performance is not the only requirement when considering the adoption of this technology to clinical practice. Trust and explainability of decisions are as important as performance. With this purpose, we also provide a method to generate visual explanations of the outcome of our polyp detector. These explanations can be used to build a physician's trust in the system and also to convey information about the inner working of the method to the designer for debugging purposes.