The evolution of biomedical imaging technology is allowing the digitization of hundreds of glass slides at once. There are multiple microscope scanners available in the market including low-cost solutions that can serve small centers. Moreover, new technology is being researched to acquire images and new modalities are appearing in the market such as electron microscopy. This reality offers new diagnostics tools to clinical practice but emphasizes also the lack of multivendor system’s interoperability. Without the adoption of standard data formats and communications methods, it will be impossible to build this industry through the installation of vendor-neutral archives and the establishment of telepathology services in the cloud. The DICOM protocol is a feasible solution to the aforementioned problem because it already provides an interface for visible light and whole slide microscope imaging modalities. While some scanners currently have DICOM interfaces, the vast majority of manufacturers continue to use proprietary solutions. This article proposes an automated DICOMization pipeline that can efficiently transform distinct proprietary microscope images from CLSM, FIB-SEM, and WSI scanners into standard DICOM with their biological information maintained within their metadata. The system feasibility and performance were evaluated with fifteen distinct proprietary modalities, including stacked WSI samples. The results demonstrated that the proposed methodology is accurate and can be used in production. The normalized objects were stored through the standard communications in the Dicoogle open-source archive.
Advances in microscopy hardware and storage capabilities lead to increasingly larger multidimensional datasets. The multiple dimensions are commonly associated with space, time, and color channels. Since “seeing is believing”, it is important to have easy access to user-friendly visualization software. Here we present IMAGE-IN, an interactive web-based multidimensional (N-D) viewer designed specifically for confocal laser scanning microscopy (CLSM) and focused ion beam scanning electron microscopy (FIB-SEM) data, with the goal of assisting biologists in their visualization and analysis tasks and promoting digital workflows. This new visualization platform includes intuitive multidimensional opacity fine-tuning, shading on/off, multiple blending modes for volume viewers, and the ability to handle multichannel volumetric data in volume and surface views. The software accepts a sequence of image files or stacked 3D images as input and offers a variety of viewing options ranging from 3D volume/surface rendering to multiplanar reconstruction approaches. We evaluate the performance by comparing the loading and rendering timings of a heterogeneous dataset of multichannel CLSM and FIB-SEM images on two devices with installed graphic cards, as well as comparing rendered image quality between ClearVolume (the ImageJ open-source desktop viewer), Napari (the Python desktop viewer), Imaris (the closed-source desktop viewer), and our proposed IMAGE-IN web viewer.
As digital medical imaging becomes more prevalent and archives increase in size, representation learning exposes an interesting opportunity for enhanced medical decision support systems. On the other hand, medical imaging data is often scarce and short on annotations. In this paper, we present an assessment of unsupervised feature learning approaches for images in biomedical literature which can be applied to automatic biomedical concept detection. Six unsupervised representation learning methods were built, including traditional bags of visual words, autoencoders, and generative adversarial networks. Each model was trained, and their respective feature spaces evaluated using images from the ImageCLEF 2017 concept detection task. The highest mean score of was obtained using representations from an adversarial autoencoder, which increased to when combined with the representations from the sparse denoising autoencoder. We conclude that it is possible to obtain more powerful representations with modern deep learning approaches than with previously popular computer vision methods. The possibility of semi-supervised learning as well as its use in medical information retrieval problems are the next steps to be strongly considered.
Sources that generate symbolic sequences with algorithmic nature may differ in statistical complexity because they create structures that follow algorithmic schemes, rather than generating symbols from a probabilistic function assuming independence. In the case of Turing machines, this means that machines with the same algorithmic complexity can create tapes with different statistical complexity. In this paper, we use a compression-based approach to measure global and local statistical complexity of specific Turing machine tapes with the same number of states and alphabet. Both measures are estimated using the best-order Markov model. For the global measure, we use the Normalized Compression (NC), while, for the local measures, we define and use normal and dynamic complexity profiles to quantify and localize lower and higher regions of statistical complexity. We assessed the validity of our methodology on synthetic and real genomic data showing that it is tolerant to increasing rates of editions and block permutations. Regarding the analysis of the tapes, we localize patterns of higher statistical complexity in two regions, for a different number of machine states. We show that these patterns are generated by a decrease of the tape’s amplitude, given the setting of small rule cycles. Additionally, we performed a comparison with a measure that uses both algorithmic and statistical approaches (BDM) for analysis of the tapes. Naturally, BDM is efficient given the algorithmic nature of the tapes. However, for a higher number of states, BDM is progressively approximated by our methodology. Finally, we provide a simple algorithm to increase the statistical complexity of a Turing machine tape while retaining the same algorithmic complexity. We supply a publicly available implementation of the algorithm in C++ language under the GPLv3 license. All results can be reproduced in full with scripts provided at the repository.
Mitochondria are intracellular organelles that act as powerhouses by breaking down nutrition molecules to produce adenosine triphosphate (ATP) as cellular fuel. They have their own genetic material called mitochondrial DNA. Alterations in mitochondrial DNA can result in primary mitochondrial diseases, including neurodegenerative disorders. Early detection of these abnormalities is crucial in slowing disease progression. With recent advances in data acquisition techniques such as focused ion beam scanning electron microscopy, it has become feasible to capture large intracellular organelle volumes at data rates reaching 4Tb/minute, each containing numerous cells. However, manually segmenting large data volumes (gigapixels) can be time-consuming for pathologists. Therefore, there is an urgent need for automated tools that can efficiently segment mitochondria with minimal user intervention. Our article proposes an ensemble of two automatic segmentation pipelines to predict regions of interest specific to mitochondria. This architecture combines the predicted outputs from both pipelines using an ensemble learning-based entropy-weighted fusion technique. The methodology minimizes the impact of individual predictions and enhances the overall segmentation results. The performance of the segmentation task is evaluated using various metrics, ensuring the reliability of our results. We used four publicly available datasets to evaluate our proposed method’s effectiveness. Our proposed fusion method has achieved a high score in terms of the mean Jaccard index and dice coefficient for all four datasets. For instance, in the UroCell dataset, our proposed fusion method achieved scores of 0.9644 for the mean Jaccard index and 0.9749 for the Dice coefficient. The mean error rate and pixel accuracy were 0.0062 and 0.9938, respectively. Later, we compared it with state-of-the-art methods like 2D and 3D CNN algorithms. Our ensemble approach shows promising segmentation efficiency with minimal intervention and can potentially aid in the early detection and mitigation of mitochondrial diseases.