Automatic Region Of Interest Extraction Research Articles

Direct Gene Expression Profile Prediction for Uveal Melanoma from Digital Cytopathology Images via Deep Learning and Salient Image Region Identification

To demonstrate that deep learning (DL) methods can produce robust prediction of gene expression profile (GEP) in uveal melanoma (UM) based on digital cytopathology images. Evaluation of a diagnostic test or technology. Deidentified smeared cytology slides stained with hematoxylin and eosin obtained from a fine needle aspirated from UM. Digital whole-slide images were generated by fine-needle aspiration biopsies of UM tumors that underwent GEP testing. A multistage DL system was developed with automatic region-of-interest (ROI) extraction from digital cytopathology images, an attention-based neural network, ROI feature aggregation, and slide-level data augmentation. The ability of our DL system in predicting GEP on a slide (patient) level. Data were partitioned at the patient level (73% training; 27% testing). In total, our study included 89 whole-slide images from 82 patients and 121 388 unique ROIs. The testing set included 24 slides from 24 patients (12 class 1 tumors; 12 class 2 tumors; 1 slide per patient). Our DL system for GEP prediction achieved an area under the receiver operating characteristic curve of 0.944, an accuracy of 91.7%, a sensitivity of 91.7%, and a specificity of 91.7% on a slide-level analysis. The incorporation of slide-level feature aggregation and data augmentation produced a more predictive DL model (P= 0.0031). Our current work established a complete pipeline for GEP prediction in UM tumors: from automatic ROI extraction from digital cytopathology whole-slide images to slide-level predictions. Our DL system demonstrated robust performance and, if validated prospectively, could serve as an image-based alternative to GEP testing.

12-APR Segmentation and Global Hu-F Descriptor for Human Spine MRI Image Retrieval

The image retrieval system has been used to provide the needed correct images to the physicians while the diagnosis and treatment process is being conducted. The earlier image retrieval system was a text-based image retrieval system (TBIRS) that used keywords for the image context and it requires human’s help to manually make text annotation on the images. The text annotation process is a laborious task especially when dealing with a huge database and is prone to human errors. To overcome the aforementioned issues, the approach of a content-based image retrieval system (CBIRS) with automatic indexing using visual features such as colour, shape and texture becomes popular. Thus, this study proposes a semi-automated shape segmentation method using a 12-anatomical point representation method of the human spine vertebrae for CBIRS. The 12 points, which are annotated manually on the region of interest (ROI), is followed by automatic ROI extraction. The segmentation method performs excellently, as evidenced by the highest accuracy of 0.9987, specificity of 0.9989, and sensitivity of 0.9913. The features of the segmented ROI are extracted with a novel global Hu-F descriptor that combines a global shape descriptor, a Hu moment invariant, and a Fourier descriptor based on the ANOVA selection approach. The retrieval phase is implemented using 100 MRI data of the human spine for thoracic, lumbar, and sacral bones. The highest obtained precision is 0.9110 using a normalized Manhattan metric for lumbar bones. In a conclusion, a retrieval system to retrieve lumbar bones of the MRI human spine has been successfully developed to help radiologists in diagnosing human spine diseases.

Human-Perception-Oriented Pseudo Analog Video Transmissions With Deep Learning

Recently, pseudo analog transmission has gained increasing attentions due to its ability to alleviate the cliff effect in video multicast scenarios. The existing pseudo analog systems are sorely optimized under the minimum mean squared error criterion without taking the perceptual video quality into consideration. In this paper, we propose a human-perception-based pseudo analog video transmission system named ROIC-Cast, which aims to intelligently enhance the transmission quality of the region-of-interest (ROI) parts. Firstly, the classic deep learning based saliency detection algorithm is adopted to decompose the continuous video sequences into ROI and non-ROI blocks. Secondly, an effective compression method is used to reduce the data amount of side information generated by the ROI extraction module. Then, the power allocation scheme is formulated as a convex problem, and the optimal transmission power for both ROI and non-ROI blocks is derived in a closed form. Finally, the simulations are conducted to validate the proposed system by comparing with a few of existing systems, e.g., KMV-Cast, SoftCast, and DAC-RAN. The proposed ROIC-Cast can achieve over 4.1dB peak signal- to-noise ratio gains of ROI compared with other systems, given the channel signal-to-noise ratio as -5dB, 0dB, 5dB, and 10dB, respectively. This significant performance improvement is due to the automatic ROI extraction, high-efficiency data compression as well as adaptive power allocation.

Small target detection based on automatic ROI extraction and local directional gray&entropy contrast map

As a special problem in object recognition, how to detect the small target from complex background fast and precisely is kept an open topic. An excellent algorithm should have ideal Pd (Probability of detection) as well as lower Fa (False alarm) provided that no missing detection. Taking it as the goal, in this paper, we put forward an enhanced version based on LDG&ECM (Local Directional Gray&Entropy Contrast Map). To enhance the discrimination of the target with the background clutter, the saliency feature of “contrast” is remearsured by associating the “gray intensity” with “entropy”. Also, with the introduction of “directivity” in the calculation, the stubborn clutter edge with similar property to the target is removed effectively. Moreover, to speed up the algorithm, the cGANs (Conditional Generative Adversarial Networks) is first introduced to extract ROI (Region of Interest) automatically. Based on DNNs, the traditional extraction methods relying on empirical threshold is improved to an end-to-end way. Additional, to detect the multiple targets in heterogeneous background, the original image is segmented into multiple homogeneous parts by OTSU. And then, a novel local adaptive threshold decision making is designed. By this way, all of the real targets with prominent local extreme property are detected and the false alarm caused by local extreme is avoided due to the balance of global information. It is verified that the proposed automation and adaptively approach has significant performance improvement compared with the state-of-the-art algorithms in various scenarios.

Carotid wall segmentation in longitudinal ultrasound images using structured random forest

Edge detection is a primary image processing technique used for object detection, data extraction, and image segmentation. Recently, edge-based segmentation using structured classifiers has been receiving increasing attention. The intima media thickness (IMT) of the common carotid artery is mainly used as a primitive indicator for the development of cardiovascular disease. For efficient measurement of the IMT, we propose a fast edge-detection technique based on a structured random forest classifier. The accuracy of IMT measurement is degraded owing to the speckle noise found in carotid ultrasound images. To address this issue, we propose the use of a state-of-the-art denoising method to reduce the speckle noise, followed by an enhancement technique to increase the contrast. Furthermore, we present a novel approach for an automatic region of interest extraction in which a pre-trained structured random forest classifier algorithm is applied for quantifying the IMT. The proposed method exhibits IMTmean ± standard deviation of 0.66mm ± 0.14, which is closer to the ground truth value 0.67mm ± 0.15 as compared to the state-of-the-art techniques.

Region-of-Interest Coding Based on Saliency Detection and Directional Wavelet for Remote Sensing Images

With growing contradiction between the high-speed acquisition of remote sensing data and the low-speed data storage and transmission, the advantages of giving higher priority to a region of interest (ROI) in compression have become prominent. Previous research focused on ROI coding, rather than automatic ROI extraction. However, accurate ROI extraction can significantly improve coding efficiency. In this letter, we propose an automatic ROI extraction based on the improved normal directional lifting wavelet transform (LWT). Then, the compression efficiency is enhanced by a novel tangent directional LWT to reduce the signal energy of high-frequency subbands. Finally, the autogenerated ROIs are encoded by a new multibitplane alternating shift method, which supports not only arbitrarily shaped ROI coding, but also flexible adjustment of compression quality in the ROI and the background. The experimental results demonstrate that our method can effectively highlight the ROIs with well-defined boundaries, meanwhile improving the ROI coding with better visual quality.

Fully automatic ROI extraction and edge-based segmentation of radius and ulna bones from hand radiographs

Bone age is a reliable measure of person's growth and maturation of skeleton. The difference between chronological age and bone age indicates presence of endocrinological problems. The automated bone age assessment system (ABAA) based on Tanner and Whitehouse method (TW3) requires monitoring the growth of radius, ulna and short bones (phalanges) of left hand. In this paper, a detailed analysis of two bones in the bone age assessment system namely, radius and ulna is presented. We propose an automatic extraction method for the region of interest (ROI) of radius and ulna bones from a left hand radiograph (RUROI). We also propose an improved edge-based segmentation technique for those bones. Quantitative and qualitative results of the proposed segmentation technique are evaluated and compared with other state-of-the-art segmentation techniques. Medical experts have also validated the qualitative results of proposed segmentation technique. Experimental results reveal that these proposed techniques provide better segmentation accuracy as compared to the other state-of-the-art segmentation techniques.

Interactive cross and multimodal biomedical image retrieval based on automatic region-of-interest (ROI) identification and classification

In biomedical articles, authors often use annotation markers such as arrows, letters, or symbols overlaid on figures and illustrations to highlight ROIs. These annotations are then referenced and correlated with concepts in the caption text or figure citations in the article text. This association creates a bridge between the visual characteristics of important regions within an image and their semantic interpretation. Identifying these assists in extracting ROIs that are likely to be highly relevant to the discussion in the article text. The aim of this work is to perform semantic search without knowing the concept keyword or the specific name of the visual pattern or appearance. We consider the problem of cross and multimodal retrieval of images from articles which contains components of text and images. Our proposed method localizes and recognizes the annotations by utilizing a combination of rule-based and statistical image processing techniques. The image regions are then annotated for classification using biomedical concepts obtained from a glossary of imaging terms. Similar automatic ROI extraction can be applied to query images, or to interactively mark an ROI. As a result, visual characteristics of the ROIs can be mapped to text concepts and then used to search image captions. In addition, the system can toggle the search process from purely perceptual to a conceptual one (crossmodal) based on utilizing user feedback or integrate both perceptual and conceptual search in a multimodal search process. The hypothesis, that such approaches would improve biomedical image retrieval, was validated through experiments on a biomedical article dataset of thoracic CT scans.

A novel approach for automatic region of interest extraction in industrial radiographic images

The increase in dearth of image analysis and interpretation in the field of remote sensing, communication, medical imaging, forensics and industrial inspection demands for a complete automated image processing system. Especially, automating the process of region of interest extraction is a great challenge as it depends completely on the domain information. Many of the researchers had attempted to automate the process of image interpretation and classification only. In particular, automating ROI extraction process in greyscale images containing weld defects had demonstrated great difficulty towards the accuracy. This paper attempted to address the issue of complete automation of greyscale image ROI extraction for further image analysis and classification. The attempt also outlines the necessity of segregating background and foreground images in industrial radiographs through a new intensity-based filtering method. A study is made to determine an appropriate segmentation method for different defect detection. In order to evaluate the designed methodology, industrial radiographs containing defects of welding process are considered.

Automatic Extraction of Caudate and Putamen in [$^11$C] Raclopride PET Using Deformable Surface Models and Normalized Cuts

Manual region-of-interest (ROI) definition is the conventional basis for calculating regional receptor binding potential (BP) values in neuroreceptor PET studies. In this study, we propose a fully automatic method to extract the left and right caudate and putamen in [ $^11$ C]raclopride BP images. The automatic extraction of these brain structures is challenging due to the small size of the structures compared to the imaging resolution and image noise. The automation of the ROI extraction could remedy the problems relating to intra- and inter-observer variability introduced by the manual ROI drawing. We compared our method to the manual ROI drawing. The automatic ROI extraction improved test-retest reproducibility of the calculated BP values compared to the manual ROI drawing. In addition, the method was tested on a Monte Carlo simulated PET image. In this experiment, the automatic method yielded more accurate BP values than those computed based on the anatomical ground-truth and manual ROI extraction. Hence, our method for the automatic ROI extraction is reliable and it can simplify and speed up the analysis of large sets of PET scans e.g., in drug development studies.