Endoscopic image analysis in semantic space

Intensive information is emerged in social media every second. Many breaking news often appear first in social media, much earlier than they appear in traditional news media. Through the technology of event detection on social media data streams, scatter information can be gathered together to inform us the popular events discussing online. An event is often modeled as a cluster of documents which discuss the same subject. Traditional event detection methods perform poorly on social media because of their huge amount of data and irregular expressions. In this paper, we propose a simple yet efficient event detection method towards social media. An event is represented by a sequence of keywords extracted from social media. We use a single-pass incremental clustering method with a trained encoder mapping documents and events into the same semantic space, which is helpful for the similarity calculation between them. We consider the similarity calculation between a tweet and an event as a matching process and construct a relevance matching dataset with tweet-event pairs. We finetune BERT (Bidirectional Encoder Representations from Transformers) model in the matching dataset to get an appropriate semantic encoder. Keywords are dynamically changed to represent an event for capturing the development of the event. Our proposed method achieves 0.86 on NMI (Normed Mutual Information), 0.69 on ARI (Adjusted Rand Index) and 0.70 on F1-score on a public twitter dataset, which shows the superiority of our method compared with baseline methods.

In this paper, a learning method is proposed to improve the retrieval process in image databases. This method uses the search transaction logs in the system and user relevance feedback scores to create a semantic space of the image database. The semantic space includes many semantic classes and all the images in the database are clustered to these classes with different membership values. The sparsity problem in the transaction logs data is solved by filling the missing values by an estimation based on the image contents and image similarities. A Fuzzy clustering algorithm is developed to create the semantic classes and find image memberships in the classes.

Text steganography helps protect citizens' freedom of speech and privacy in cyberspace by constructing innocent-looking texts to evade censorship and surveillance. Existing methods, especially generative ones, rely on delicate character-level manipulations, making them vulnerable to failure even under minor alterations. This paper introduces a novel framework that shifts from character-level to semantic space-based information hiding, which greatly improves improving robustness and reliability. The framework comprises three phases: preparation, where a stable semantic space is designed for hiding information; embedding, where a reversible codebook maps binary messages to semantemes via semantic encoding and source coding; and synthesis, where large language models act as multi-agent systems to produce stegotexts that preserve semantic consistency. Extensive experiments show that our method matches state-of-the-art techniques in hiding capacity and text quality, while significantly surpassing them in robustness-achieving at least an 81.2% higher correct message rate under three types of attacks. These findings highlight the strong potential of semantic-based text steganography.

Computer-aided cancer diagnosis (CAD) methods based on the histopathological images have achieved great development. The content-based whole slide image (WSI) retrieval is one of the important application that can search for the informative data to assist clinical diagnosis. It is notable that the current retrieval system are mainly developed based on the image content and image labels. The diagnosis report for the WSIs given by the pathologists are also valuable data, but have not yet been adequately considered in modeling. In this paper, we propose a cross-modal retrieval framework based on histopathology WSIs and diagnosis report, which can simultaneously achieve four retrieval tasks for histopathology database across WSIs and diagnosis reports. The compact binary features from both WSIs and diagnosis reports are first extracted, and then built in a common vision-language semantic feature space by the constraint of the designed cross hashing loss function. The method was verified on a gastric histopathology dataset that contains 932 gastric cases with 4 lesion categories. Experimental results have demonstrated the effectiveness of the proposed method in the cross-modal retrieval tasks for digital pathology system.

Multi-level annotation of images is a promising solution to enable more effective semantic image retrieval by using various keywords at different semantic levels. In this paper, we propose a multi-level approach to annotate the semantics of natural scenes by using both the dominant image components and the relevant semantic concepts. In contrast to the well-known image-based and region-based approaches, we use the salient objects as the dominant image components to achieve automatic image annotation at the content level. By using the salient objects for image content representation, a novel image classification technique is developed to achieve automatic image annotation at the concept level. To detect the salient objects automatically, a set of detection functions are learned from the labeled image regions by using Support Vector Machine (SVM) classifiers with an automatic scheme for searching the optimal model parameters. To generate the semantic concepts, finite mixture models are used to approximate the class distributions of the relevant salient objects. An adaptive EM algorithm has been proposed to determine the optimal model structure and model parameters simultaneously. We have also demonstrated that our algorithms are very effective to enable multi-level annotation of natural scenes in a large-scale dataset.

Coronary Angiography: The Need for Improvement and the Barriers to Adoption of New Technology

© 2016 IEEE. Spatio-temporal cues offer a rich source of information for inferring structural and semantic scene properties. A particularly useful representation in computer vision is a spatio-temporal video segmentation. Together with motion, knowledge of depth can substantially improve superpixel segmentation. In this work we present a novel framework for spatio-temporal segmentation from RGBD video. The method employs both low-level (intensity, color) and high-level (deformable parts model) appearance features. Motion is incorporated through the use of optical flow to construct the temporal connections in the graph Laplacian. Depth cues are incorporated in the similarity metric to provide an informative cue for object boundaries at depth disparities. Naïve application of spectral clustering to dense spatio-temporal graphs leads to a high computational cost that is typically addressed through the use of GPUs or computer clusters. By contrast, we build upon a recently proposed Nyström approximation strategy for spatio-temporal clustering that enables computation on a single core. We further explore structured local connectivity patterns to give high performance at low computational cost. Also we propose a novel context-aware aggregation method that uses a deformable parts model to group the detected parts of the object as a single segment with an accurate boundary. Detailed experiments on the NYU Depth Dataset and TUM RGBD Dataset is performed to compare against previous large-scale graph-based spatio-temporal segmentation techniques which shows the substantial performance advantages of our framework.

A 9-year-old female presented to the emergency department complaining of intense peri-umbilical pain with nausea and vomiting after a road accident. Physical examination revealed abdominal tenderness with a painful well-circumscribed mass in the lower abdomen. The haemodynamic state was stable. Ultrasonography showed a mobile hypoechoic mass with a moderate amount of free intraperitoneal fluid. An abdominal CT scan was performed. Axial images before and after intravenously administered contrast material are shown. What are the abnormalities seen in Figure 1? Figure 1 Abdominal CT scans obtained before and after intravenously administered contrast material. What do you notice in Figure 2? Figure 2 Abdominal CT scan with intravenously administered contrast material at a higher level. What is your diagnosis?

Computer-Aided Diagnosis With Monochromatic Light Endoscopy for Scoring Histologic Remission in Ulcerative Colitis

The natural environment at high altitude triggers plateau diseases. This requires excellent military physicians that have graduated from high altitude military medical training to serve people on the plateau. High altitude military medical undergraduates need to complete a variety of military medicine courses. High altitude military medical geography is an important subject in military medicine that focuses on the relationship between the high altitude geographical environment and the health and medical support of military members. Considering military medical geographic information is a key aspect before military actions. This article summarizes the experience of teaching a high altitude military medical geography program, which has produced Chinese military medical undergraduates with excellent theoretical knowledge and practical skills. Our teaching process includes the first and second class levels. The former comprises theoretical and practical courses, including investigations in the field, practice on the plateau, and practice with geographic information systems. During the first class level teaching process, theoretical knowledge was reorganized and combined into four modules. After moving to the second class level, which last for 2–3 years, students continue to develop their military medical geography expertise under the guidance of teachers in summer school. This characteristic teaching mode supports the training of excellent high altitude military physicians for China.

Introduction: Warts (verrucae) are benign epidermal proliferations caused by Human Papillomavirus (HPV), with a global prevalence estimated at 7-12% in the general population and higher incidence in children and immunocompromised individuals. Clinical diagnosis is usually straightforward; however, dermoscopy provides enhanced visualization of vascular structures and surface changes, aiding in differentiation from calluses, corns and seborrheic keratosis. Recently, Artificial Intelligence (AI)-based diagnostic tools have shown promise in dermatology, particularly in image-based classification. This study explores the role of AI in dermoscopic diagnosis of warts. Methods: Dermoscopic images of clinically confirmed warts were collected and compared with images of clinically similar benign lesions. The AI model was trained to identify key parameters including vascular morphology (dotted, linear or looped capillaries), papillomatous surface, hemorrhagic dots and keratin distribution. Images were split into training, validation and testing cohorts. Diagnostic metrics included accuracy, sensitivity, specificity and Area Under the ROC Curve (AUC). Results: Dermoscopic features most predictive of warts included densely packed dotted and linear vessels surrounded by whitish halos, papillomatous projections and thrombosed capillaries appearing as black or red dots. Compared to human dermoscopic interpretation, the AI system demonstrated higher diagnostic consistency and reduced observer variability, with improved overall diagnostic performance. Conclusion: Incorporating AI into dermoscopic evaluation of warts may improve diagnostic accuracy, reduce unnecessary treatments and enhance clinical decision-making. Larger multicentric datasets and prospective validation are required to optimize AI integration into routine dermatology practice.

Increasing quality and performance of artificial intelligence (AI) in general and machine learning (ML) in particular is followed by a wider use of these approaches in everyday life. As part of this development, ML classifiers have also gained more importance for diagnosing diseases within biomedical engineering and medical sciences. However, many of those ubiquitous high-performing ML algorithms reveal a black-box-nature, leading to opaque and incomprehensible systems that complicate human interpretations of single predictions or the whole prediction process. This puts up a serious challenge on human decision makers to develop trust, which is much needed in life-changing decision tasks. This paper is designed to answer the question how expert companion systems for decision support can be designed to be interpretable and therefore transparent and comprehensible for humans. On the other hand, an approach for interactive ML as well as human-in-the-loop-learning is demonstrated in order to integrate human expert knowledge into ML models so that humans and machines act as companions within a critical decision task. We especially address the problem of Semantic Alignment between ML classifiers and its human users as a prerequisite for semantically relevant and useful explanations as well as interactions. Our roadmap paper presents and discusses an interdisciplinary yet integrated Comprehensible Artificial Intelligence (cAI)-transition-framework with regard to the task of medical diagnosis. We explain and integrate relevant concepts and research areas to provide the reader with a hands-on-cookbook for achieving the transition from opaque black-box models to interactive, transparent, comprehensible and trustworthy systems. To make our approach tangible, we present suitable state of the art methods with regard to the medical domain and include a realization concept of our framework. The emphasis is on the concept of Mutual Explanations (ME) that we introduce as a dialog-based, incremental process in order to provide human ML users with trust, but also with stronger participation within the learning process.

Traditional image retrieval depends on the images embedded in text messages, text description of the limitations of image content, resulting in low quality of image retrieval. The local information extracted image itself, the use of local features LSH image matching algorithm, memory requirements has led to a linear growth. To overcome these shortcomings, then propose the method of image retrieval which based on the network salient features, by screening out the salient features of the frequent appearance of points, with each image feature point matching vectors to generate histogram matching, the use of its similarity to generate the calculation of PageRank, Adjacency matrix for each image to generate a PageRank score. Experimental results show that the algorithm improves the network image retrieval efficiency and accuracy.

Reality, Teaching, and the Four Researchers

The Indian court system is dominated by English, the language is spoken by less than ten percent of the population. Such a linguistic imbalance places significant barriers for citizens wishing to meaningfully understand or even receive information about the law, which is an area of highly technical employment of periodical judgments and dense conceptualisation of vocabulary. The state of Legal NLP research in India, while being richly multilingual, has seen a predominance of research concerning English, with most of the work within the country on multilingual challenges or inclusivity issues. This paper aims at bridging this accessibility gap by presenting a multilingual semantic legal information retrieval framework that allows the users to access the semantically relevant legal cases in twenty-two Indian regional languages. Based on the bidirectional neural translation model, IndicTrans2 and a transformer-based semantic representation embedding model, the proposed system can encode legal texts into the common English embedding semantic space. Language neutral queries posed in local languages are translated and embedded and the resulting vectors are semantically matched through HNSW-based vector similarity searching, thus guaranteeing efficient cross language retrieval. The framework was evaluated on a manually curated dataset taken from the LeSICiN corpus consisting of hundred queries in several languages with relevance-ranked case documents. Experimental results show good retrieval effects, such as Precision@1 of 0.9833 and NDCG@3 of 0.6203, which are much better than traditional lexical retrieval systems, such as BM25. The system combines semantic encodings with translation-based alignment, allowing retrieval of cases correctly even regardless of the language of the query. This is an attempt towards democratizing access to legal information, linguistic inclusivity and towards ensuring equitable access to justice in the multilingual context of the Indian judiciary.