Automatic figure ranking and user interfacing for intelligent figure search

An automatic system for extracting figure-caption pair from medical documents: a six-fold approach

Background

Figures and captions in medical documentation contain important information. As a result, researchers are becoming more interested in obtaining published medical figures from medical papers and utilizing the captions as a knowledge source.

Methods

This work introduces a unique and successful six-fold methodology for extracting figure-caption pairs. The A-torus wavelet transform is used to retrieve the first edge from the scanned page. Then, using the maximally stable extremal regions connected component feature, text and graphical contents are isolated from the edge document, and multi-layer perceptron is used to successfully detect and retrieve figures and captions from medical records. The figure-caption pair is then extracted using the bounding box approach. The files that contain the figures and captions are saved separately and supplied to the end useras theoutput of any investigation. The proposed approach is evaluated using a self-created database based on the pages collected from five open access books: Sergey Makarov, Gregory Noetscher and Aapo Nummenmaa's book "Brain and Human Body Modelling 2021", "Healthcare and Disease Burden in Africa" by Ilha Niohuru, "All-Optical Methods to Study Neuronal Function" by Eirini Papagiakoumou, "RNA, the Epicenter of Genetic Information" by John Mattick and Paulo Amaral and "Illustrated Manual of Pediatric Dermatology" by Susan Bayliss Mallory, Alanna Bree and Peggy Chern.

Results

Experiments and findings comparing the new method to earlier systems reveal a significant increase in efficiency, demonstrating the suggested technique's robustness and efficiency.

Visual Summary Identification From Scientific Publications via Self-Supervised Learning

The exponential growth of scientific literature yields the need to support users to both effectively and efficiently analyze and understand the some body of research work. This exploratory process can be facilitated by providing graphical abstracts–a visual summary of a scientific publication. Accordingly, previous work recently presented an initial study on automatic identification of a central figure in a scientific publication, to be used as the publication’s visual summary. This study, however, have been limited only to a single (biomedical) domain. This is primarily because the current state-of-the-art relies on supervised machine learning, typically relying on the existence of large amounts of labeled data: the only existing annotated data set until now covered only the biomedical publications. In this work, we build a novel benchmark data set for visual summary identification from scientific publications, which consists of papers presented at conferences from several areas of computer science. We couple this contribution with a new self-supervised learning approach to learn a heuristic matching of in-text references to figures with figure captions. Our self-supervised pre-training, executed on a large unlabeled collection of publications, attenuates the need for large annotated data sets for visual summary identification and facilitates domain transfer for this task. We evaluate our self-supervised pretraining for visual summary identification on both the existing biomedical and our newly presented computer science data set. The experimental results suggest that the proposed method is able to outperform the previous state-of-the-art without any task-specific annotations.

Figure and caption extraction from biomedical documents

Motivation

Figures and captions convey essential information in biomedical documents. As such, there is a growing interest in mining published biomedical figures and in utilizing their respective captions as a source of knowledge. Notably, an essential step underlying such mining is the extraction of figures and captions from publications. While several PDF parsing tools that extract information from such documents are publicly available, they attempt to identify images by analyzing the PDF encoding and structure and the complex graphical objects embedded within. As such, they often incorrectly identify figures and captions in scientific publications, whose structure is often non-trivial. The extraction of figures, captions and figure-caption pairs from biomedical publications is thus neither well-studied nor yet well-addressed.

Results

We introduce a new and effective system for figure and caption extraction, PDFigCapX. Unlike existing methods, we first separate between text and graphical contents, and then utilize layout information to effectively detect and extract figures and captions. We generate files containing the figures and their associated captions and provide those as output to the end-user.

We test our system both over a public dataset of computer science documents previously used by others, and over two newly collected sets of publications focusing on the biomedical domain. Our experiments and results comparing PDFigCapX to other state-of-the-art systems show a significant improvement in performance, and demonstrate the effectiveness and robustness of our approach.

Availability and implementation

Our system is publicly available for use at: https://www.eecis.udel.edu/~compbio/PDFigCapX. The two new datasets are available at: https://www.eecis.udel.edu/~compbio/PDFigCapX/Downloads

Large expert-curated database for benchmarking document similarity detection in biomedical literature search

Document recommendation systems for locating relevant literature have mostly relied on methods developed a decade ago. This is largely due to the lack of a large offline gold-standard benchmark of relevant documents that cover a variety of research fields such that newly developed literature search techniques can be compared, improved and translated into practice. To overcome this bottleneck, we have established the RElevant LIterature SearcH consortium consisting of more than 1500 scientists from 84 countries, who have collectively annotated the relevance of over 180 000 PubMed-listed articles with regard to their respective seed (input) article/s. The majority of annotations were contributed by highly experienced, original authors of the seed articles. The collected data cover 76% of all unique PubMed Medical Subject Headings descriptors. No systematic biases were observed across different experience levels, research fields or time spent on annotations. More importantly, annotations of the same document pairs contributed by different scientists were highly concordant. We further show that the three representative baseline methods used to generate recommended articles for evaluation (Okapi Best Matching 25, Term Frequency-Inverse Document Frequency and PubMed Related Articles) had similar overall performances. Additionally, we found that these methods each tend to produce distinct collections of recommended articles, suggesting that a hybrid method may be required to completely capture all relevant articles. The established database server located at https://relishdb.ict.griffith.edu.au is freely available for the downloading of annotation data and the blind testing of new methods. We expect that this benchmark will be useful for stimulating the development of new powerful techniques for title and title/abstract-based search engines for relevant articles in biomedical research.

Large expert-curated database for benchmarking document similarity detection in biomedical literature search

Document recommendation systems for locating relevant literature have mostly relied on methods developed a decade ago. This is largely due to the lack of a large offline gold-standard benchmark of relevant documents that cover a variety of research fields such that newly developed literature search techniques can be compared, improved and translated into practice. To overcome this bottleneck, we have established the RElevant LIterature SearcH consortium consisting of more than 1500 scientists from 84 countries, who have collectively annotated the relevance of over 180 000 PubMed-listed articles with regard to their respective seed (input) article/s. The majority of annotations were contributed by highly experienced, original authors of the seed articles. The collected data cover 76% of all unique PubMed Medical Subject Headings descriptors. No systematic biases were observed across different experience levels, research fields or time spent on annotations. More importantly, annotations of the same document pairs contributed by different scientists were highly concordant. We further show that the three representative baseline methods used to generate recommended articles for evaluation (Okapi Best Matching 25, Term Frequency-Inverse Document Frequency and PubMed Related Articles) had similar overall performances. Additionally, we found that these methods each tend to produce distinct collections of recommended articles, suggesting that a hybrid method may be required to completely capture all relevant articles. The established database server located at https://relishdb.ict.griffith.edu.au is freely available for the downloading of annotation data and the blind testing of new methods. We expect that this benchmark will be useful for stimulating the development of new powerful techniques for title and title/abstract-based search engines for relevant articles in biomedical research.

DeTEXT: A Database for Evaluating Text Extraction from Biomedical Literature Figures

Hundreds of millions of figures are available in biomedical literature, representing important biomedical experimental evidence. Since text is a rich source of information in figures, automatically extracting such text may assist in the task of mining figure information. A high-quality ground truth standard can greatly facilitate the development of an automated system. This article describes DeTEXT: A database for evaluating text extraction from biomedical literature figures. It is the first publicly available, human-annotated, high quality, and large-scale figure-text dataset with 288 full-text articles, 500 biomedical figures, and 9308 text regions. This article describes how figures were selected from open-access full-text biomedical articles and how annotation guidelines and annotation tools were developed. We also discuss the inter-annotator agreement and the reliability of the annotations. We summarize the statistics of the DeTEXT data and make available evaluation protocols for DeTEXT. Finally we lay out challenges we observed in the automated detection and recognition of figure text and discuss research directions in this area. DeTEXT is publicly available for downloading at http://prir.ustb.edu.cn/DeTEXT/.

Learning to rank figures within a biomedical article

Hundreds of millions of figures are available in biomedical literature, representing important biomedical experimental evidence. This ever-increasing sheer volume has made it difficult for scientists to effectively and accurately access figures of their interest, the process of which is crucial for validating research facts and for formulating or testing novel research hypotheses. Current figure search applications can't fully meet this challenge as the “bag of figures” assumption doesn't take into account the relationship among figures. In our previous study, hundreds of biomedical researchers have annotated articles in which they serve as corresponding authors. They ranked each figure in their paper based on a figure's importance at their discretion, referred to as “figure ranking”. Using this collection of annotated data, we investigated computational approaches to automatically rank figures. We exploited and extended the state-of-the-art listwise learning-to-rank algorithms and developed a new supervised-learning model BioFigRank.

The cross-validation results show that BioFigRank yielded the best performance compared with other state-of-the-art computational models, and the greedy feature selection can further boost the ranking performance significantly. Furthermore, we carry out the evaluation by comparing BioFigRank with three-level competitive domain-specific human experts: (1) First Author, (2) Non-Author-In-Domain-Expert who is not the author nor co-author of an article but who works in the same field of the corresponding author of the article, and (3) Non-Author-Out-Domain-Expert who is not the author nor co-author of an article and who may or may not work in the same field of the corresponding author of an article. Our results show that BioFigRank outperforms Non-Author-Out-Domain-Expert and performs as well as Non-Author-In-Domain-Expert. Although BioFigRank underperforms First Author, since most biomedical researchers are either in- or out-domain-experts for an article, we conclude that BioFigRank represents an artificial intelligence system that offers expert-level intelligence to help biomedical researchers to navigate increasingly proliferated big data efficiently.

Beyond captions: linking figures with abstract sentences in biomedical articles

Although figures in scientific articles have high information content and concisely communicate many key research findings, they are currently under utilized by literature search and retrieval systems. Many systems ignore figures, and those that do not typically only consider caption text. This study describes and evaluates a fully automated approach for associating figures in the body of a biomedical article with sentences in its abstract. We use supervised methods to learn probabilistic language models, hidden Markov models, and conditional random fields for predicting associations between abstract sentences and figures. Three kinds of evidence are used: text in abstract sentences and figures, relative positions of sentences and figures, and the patterns of sentence/figure associations across an article. Each information source is shown to have predictive value, and models that use all kinds of evidence are more accurate than models that do not. Our most accurate method has an -score of 69% on a cross-validation experiment, is competitive with the accuracy of human experts, has significantly better predictive accuracy than state-of-the-art methods and enables users to access figures associated with an abstract sentence with an average of 1.82 fewer mouse clicks. A user evaluation shows that human users find our system beneficial. The system is available at http://FigureItOut.askHERMES.org.

Automatic figure classification in bioscience literature

Millions of figures appear in biomedical articles, and it is important to develop an intelligent figure search engine to return relevant figures based on user entries. In this study we report a figure classifier that automatically classifies biomedical figures into five predefined figure types: Gel-image, Image-of-thing, Graph, Model, and Mix. The classifier explored rich image features and integrated them with text features. We performed feature selection and explored different classification models, including a rule-based figure classifier, a supervised machine-learning classifier, and a multi-model classifier, the latter of which integrated the first two classifiers. Our results show that feature selection improved figure classification and the novel image features we explored were the best among image features that we have examined. Our results also show that integrating text and image features achieved better performance than using either of them individually. The best system is a multi-model classifier which combines the rule-based hierarchical classifier and a support vector machine (SVM) based classifier, achieving a 76.7% F1-score for five-type classification. We demonstrated our system at http://figureclassification.askhermes.org/.

The biomedical discourse relation bank

Background

Identification of discourse relations, such as causal and contrastive relations, between situations mentioned in text is an important task for biomedical text-mining. A biomedical text corpus annotated with discourse relations would be very useful for developing and evaluating methods for biomedical discourse processing. However, little effort has been made to develop such an annotated resource.

Results

We have developed the Biomedical Discourse Relation Bank (BioDRB), in which we have annotated explicit and implicit discourse relations in 24 open-access full-text biomedical articles from the GENIA corpus. Guidelines for the annotation were adapted from the Penn Discourse TreeBank (PDTB), which has discourse relations annotated over open-domain news articles. We introduced new conventions and modifications to the sense classification. We report reliable inter-annotator agreement of over 80% for all sub-tasks. Experiments for identifying the sense of explicit discourse connectives show the connective itself as a highly reliable indicator for coarse sense classification (accuracy 90.9% and F1 score 0.89). These results are comparable to results obtained with the same classifier on the PDTB data. With more refined sense classification, there is degradation in performance (accuracy 69.2% and F1 score 0.28), mainly due to sparsity in the data. The size of the corpus was found to be sufficient for identifying the sense of explicit connectives, with classifier performance stabilizing at about 1900 training instances. Finally, the classifier performs poorly when trained on PDTB and tested on BioDRB (accuracy 54.5% and F1 score 0.57).

Conclusion

Our work shows that discourse relations can be reliably annotated in biomedical text. Coarse sense disambiguation of explicit connectives can be done with high reliability by using just the connective as a feature, but more refined sense classification requires either richer features or more annotated data. The poor performance of a classifier trained in the open domain and tested in the biomedical domain suggests significant differences in the semantic usage of connectives across these domains, and provides robust evidence for a biomedical sublanguage for discourse and the need to develop a specialized biomedical discourse annotated corpus. The results of our cross-domain experiments are consistent with related work on identifying connectives in BioDRB.

Figure text extraction in biomedical literature

BACKGROUND

Figures are ubiquitous in biomedical full-text articles, and they represent important biomedical knowledge. However, the sheer volume of biomedical publications has made it necessary to develop computational approaches for accessing figures. Therefore, we are developing the Biomedical Figure Search engine (http://figuresearch.askHERMES.org) to allow bioscientists to access figures efficiently. Since text frequently appears in figures, automatically extracting such text may assist the task of mining information from figures. Little research, however, has been conducted exploring text extraction from biomedical figures.

METHODOLOGY

We first evaluated an off-the-shelf Optical Character Recognition (OCR) tool on its ability to extract text from figures appearing in biomedical full-text articles. We then developed a Figure Text Extraction Tool (FigTExT) to improve the performance of the OCR tool for figure text extraction through the use of three innovative components: image preprocessing, character recognition, and text correction. We first developed image preprocessing to enhance image quality and to improve text localization. Then we adapted the off-the-shelf OCR tool on the improved text localization for character recognition. Finally, we developed and evaluated a novel text correction framework by taking advantage of figure-specific lexicons.

RESULTS/CONCLUSIONS

The evaluation on 382 figures (9,643 figure texts in total) randomly selected from PubMed Central full-text articles shows that FigTExT performed with 84% precision, 98% recall, and 90% F1-score for text localization and with 62.5% precision, 51.0% recall and 56.2% F1-score for figure text extraction. When limiting figure texts to those judged by domain experts to be important content, FigTExT performed with 87.3% precision, 68.8% recall, and 77% F1-score. FigTExT significantly improved the performance of the off-the-shelf OCR tool we used, which on its own performed with 36.6% precision, 19.3% recall, and 25.3% F1-score for text extraction. In addition, our results show that FigTExT can extract texts that do not appear in figure captions or other associated text, further suggesting the potential utility of FigTExT for improving figure search.