Automated tissue characterization is one of the most crucial components a computer aided diagnosis (CAD) system for interstitial lung diseases (ILDs). Although much research has been conducted in this field, problem remains challenging. Deep learning techniques have recently achieved impressive results variety vision problems, raising expectations that they might be applied other domains, such as medical image analysis. In paper, we propose and evaluate convolutional neural network (CNN),...

Early diagnosis of interstitial lung diseases is crucial for their treatment, but even experienced physicians find it difficult, as clinical manifestations are similar. In order to assist with the diagnosis, computer-aided (CAD) systems have been developed. These commonly rely on a fixed scale classifier that scans CT images, recognizes textural patterns and generates map pathologies. previous study, we proposed method classifying tissue using deep convolutional neural network (CNN), an...

Computer vision-based food recognition could be used to estimate a meal's carbohydrate content for diabetic patients. This study proposes methodology automatic recognition, based on the bag-of-features (BoF) model. An extensive technical investigation was conducted identification and optimization of best performing components involved in BoF architecture, as well estimation corresponding parameters. For design evaluation prototype system, visual dataset with nearly 5000 images created...

Objectives The objective of this study is to assess the performance a computer-aided diagnosis (CAD) system (INTACT system) for automatic classification high-resolution computed tomography images into 4 radiological diagnostic categories and compare with radiologists on same task. Materials Methods For comparison, total 105 cases pulmonary fibrosis were studied (54 nonspecific interstitial pneumonia 51 usual pneumonia). All diagnoses lung disease board consensus (radiologically or...

The increasing prevalence of diet-related chronic diseases coupled with the ineffectiveness traditional diet management methods have resulted in a need for novel tools to accurately and automatically assess meals. Recently, computer vision-based systems that use meal images their content been proposed. Food portion estimation is most difficult task individuals assessing meals it also least studied area. This paper proposes three-stage system calculate sizes using two dish acquired by mobile...

Background: Diabetes mellitus is spreading throughout the world and diabetic individuals have been shown to often assess their food intake inaccurately; therefore, it a matter of urgency develop automated diet assessment tools. The recent availability mobile phones with enhanced capabilities, together advances in computer vision, permitted development image analysis apps for meals. GoCARB phone-based system designed support type 1 diabetes during daily carbohydrate estimation. In typical...

Early and accurate diagnosis of interstitial lung diseases (ILDs) is crucial for making treatment decisions, but can be challenging even experienced radiologists. The diagnostic procedure based on the detection recognition different ILD pathologies in thoracic CT scans, yet their manifestation often appears similar. In this study, we propose use a deep purely convolutional neural network semantic segmentation patterns, as basic component computer aided system ILDs. proposed CNN, which...

Individuals with type 1 diabetes (T1D) have to count the carbohydrates (CHOs) of their meal estimate prandial insulin dose needed compensate for meal's effect on blood glucose levels. CHO counting is very challenging but also crucial, since an error 20 grams can substantially impair postprandial control.The GoCARB system a smartphone application designed support T1D patients nonpacked foods. In typical scenario, user places reference card next dish and acquires 2 images his/her smartphone....

In this paper, we propose novel methodologies for the automatic segmentation and recognition of multi-food images. The proposed methods implement first modules a carbohydrate counting insulin advisory system type 1 diabetic patients. Initially plate is segmented using pyramidal mean-shift filtering region growing algorithm. Then each resulted segments described by both color texture features classified support vector machine into one six different major food classes. Finally, modified...

The prevalence of diet-related chronic diseases strongly impacts global health and services. Currently, it takes training strong personal involvement to manage or treat these diseases. One way assist with dietary assessment is through computer vision systems that can recognize foods their portion sizes from images output the corresponding nutritional information. When multiple food items may exist, a segmentation stage should also be applied before recognition. In this study, we propose...

Over the last decade, a plethora of computer-aided diagnosis (CAD) systems have been proposed aiming to improve accuracy physicians in interstitial lung diseases (ILD). In this study, we propose scheme for classification HRCT image patches with ILD abnormalities as basic component towards quantification various patterns lung. The feature extraction method relies on local spectral analysis using DCT-based filter bank. After convolving bank, q-quantiles are computed describing distribution...

Key role in the prevention of diet-related chronic diseases plays balanced nutrition together with a proper diet. The conventional dietary assessment methods are time-consuming, expensive and prone to errors. New technology-based that provide reliable convenient assessment, have emerged during last decade. advances field computer vision permitted use meal image assess nutrient content usually through three steps: food segmentation, recognition volume estimation. In this paper, we propose one...

This paper proposes a hybrid system for text detection in video frames. The consists of two main stages. In the first stage regions are detected based on edge map image leading high recall rate with minimum computation requirements. sequel, refinement uses an SVM classifier trained features obtained by new local binary pattern operator which results diminishing false alarms. Experimental show overall performance that proves discriminating ability proposed feature set.

Background: In an artificial pancreas (AP), the meals are either manually announced or detected and their size estimated from blood glucose level. Both methods have limitations, which result in suboptimal postprandial control. The GoCARB system is designed to provide carbohydrate content of presented within AP framework. Method: combined use with a control algorithm assessed series 12 computer simulations. simulations defined according type (open closed loop), not-use diabetics’ skills...

This paper proposes an algorithm for detecting artificial text in video frames using edge information. First, map is created the Canny detector. Then, morphological dilation and opening are used order to connect vertical edges eliminate false alarms. Bounding boxes determined every non-zero valued connected component, consisting initial candidate areas. Finally, projection analysis applied, refining result splitting areas lines. The whole applied different resolutions ensure detection with...

Point-of-care testing (POCT) has transformed the healthcare landscape by delivering quick and cheap diagnostic services closer to patient. Urine test strips are one of most commonly used POCT tools, however their manual interpretation can be challenging, particularly for elderlies people with eye disorders. In this study, we propose a smartphone application designed automatically perform semi-quantitative colorimetric analysis on urine using just image strip, placed specially reference card....

Dietary and lifestyle management rely on objective accurate diet assessment. To assess dietary intake itself requires training skills however, in that regard, trained individuals often misjudge what they eat, even when are under strict constraints [1]. These issues emphasize the need for objective, assessment tools can be delivered to directly used by public monitor their intake.

This paper proposes a performance evaluation method for text detection in color images. The method, contrary to previous approaches, is not based on the inexplicitly defined bounding boxes of result but considers only pixels detected by binarizing image and applying inversion if needed. Moreover, order gain independence from chosen binarization algorithm, uses skeleton binarized image. results produced proposed protocol proved be quite representative reasonable compared corresponding optical result.