There is a growing concern about chronic diseases and other health problems related to diet including obesity cancer. The need accurately measure (what foods person consumes) becomes imperative. Dietary intake provides valuable insights for mounting intervention programs prevention of diseases. Measuring accurate dietary considered be an open research problem in the nutrition fields. In this paper, we describe novel mobile telephone food record that will provide account daily nutrient...

Background: The development of a mobile telephone food record has the potential to ameliorate much burden associated with current methods dietary assessment. When using record, respondents capture an image their foods and beverages before after eating. Methods analysis volume estimation allow for automatic identification foods. To obtain suitable image, all fiducial marker must be included in image. Objective: evaluate defined set skills among adolescents adults when images compare...

We propose a method for dietary assessment to automatically identify and locate food in variety of images captured during controlled natural eating events. Two concepts are combined achieve this: set segmented objects can be partitioned into perceptually similar object classes based on global local features; used assess the accuracy image segmentation. These ideas implemented by generating multiple segmentations an select stable classifier's confidence score assigned each region. Automatic...

Modern deep learning approaches have achieved great success in many vision applications by training a model using all available task-specific data. However, there are two major obstacles making it challenging to implement for real life applications: (1) Learning new classes makes the trained quickly forget old knowledge, which is referred as catastrophic forgetting. (2) As observations of come sequentially over time, distribution may change unforeseen way, performance degrade dramatically on...

Altered and manipulated multimedia is increasingly present widely distributed via social media platforms. Advanced video manipulation tools enable the generation of highly realistic-looking altered multimedia. While many methods have been presented to detect manipulations, most them fail when evaluated with data outside datasets used in research environments. In order address this problem, Deepfake Detection Challenge (DFDC) provides a large dataset videos containing realistic manipulations...

Obtaining accurate food portion estimation automatically is challenging since the processes of preparation and consumption impose large variations on shapes appearances. The aim this paper was to estimate energy numeric value from eating occasion images captured using mobile record. To model characteristics distribution in an scene, a new concept "food distribution" introduced. mapping image its learned Generative Adversarial Network (GAN) architecture. Food estimated based predicted by GAN....

Recent work has shown a strong theoretical connection between variational autoencoders (VAEs) and the rate distortion theory. Motivated by this, we consider problem of lossy image compression from perspective generative modeling. Starting ResNet VAEs, which are originally designed for data (image) distribution modeling, redesign their latent variable model using quantization-aware posterior prior, enabling easy quantization entropy coding compression. Along with improved neural network...

This paper addresses the problem of lossy image compression, a fundamental in processing and information theory that is involved many real-world applications. We start by reviewing framework variational autoencoders (VAEs), powerful class generative probabilistic models has deep connection to compression. Based on VAEs, we develop new scheme for which name quantization-aware ResNet VAE (QARV). Our method incorporates hierarchical architecture integrated with test-time quantization training,...

Many chronic diseases, such as heart diabetes, and obesity, can be related to diet. Hence, the need accurately measure diet becomes imperative. We are developing methods use image analysis tools for identification quantification of food consumed at a meal. In this paper we describe new approach using several features based on local global measures "voting" late decision fusion classifier identify items. Experimental results wide variety items presented.

As obesity concerns mount, dietary assessment methods for prevention and intervention are being developed. These include recording, cataloging analyzing daily records to monitor energy nutrient intakes. Given the ubiquity of mobile devices with built-in cameras, one possible means improving is through photographing foods inputting these images into a system that can determine content in images. One critical issues such image-based tool accurate consistent estimation food portion sizes. The...

The mobile Food Record (mFR) is an image-based dietary assessment method for devices. study primary aim was to test the accuracy of mFR by comparing reported energy intake (rEI) total expenditure (TEE) using doubly labeled water (DLW) method. Usability assessed questionnaires before and after study. Participants were 45 community dwelling men women, 21–65 years. They provided pack-out meals snacks encouraged supplement with usual foods beverages not provided. After being dosed DLW,...

In this paper we present a food portion estimation technique based on single-view image used for the of amount energy (in kilocalories) consumed at meal. Unlike previous methods have developed, new is capable estimating without manual tuning parameters. Although 3D scene reconstruction in general an ill-posed problem, use geometric models such as shape container can help to partially recover parameters items scene. Based estimated each item and reference object scene, volume be determined....

Current satellite imaging technology enables shooting highresolution pictures of the ground. As any other kind digital images, overhead can also be easily forged. However, common image forensic techniques are often developed for consumer camera which strongly differ in their nature from ones (e.g., compression schemes, post-processing, sensors, etc.). Therefore, many accurate state-of-the-art algorithms bound to fail if blindly applied analysis. Development novel tools images is paramount...

Dietary intake provides valuable insights for mounting intervention programs prevention of disease. With growing concern adolescent obesity, the need to accurately measure diet becomes imperative. Assessment among adolescents is problematic as this group has irregular eating patterns and have less enthusiasm recording food intake. Preliminary studies suggest that innovative use technology may improve accuracy information from young people. In paper, we propose a novel record method using...

There is a growing concern about chronic diseases and other health problems related to diet including obesity cancer. Dietary intake provides valuable insights for mounting intervention programs prevention of diseases. Measuring accurate dietary considered be an open research problem in the nutritions fields. In this paper, we describe novel mobile telephone food record that measure daily nutrient intake. Our approach includes use image analysis tools identification quantification consumed...

Due to the growing concern of chronic diseases and other health problems related diet, there is a need develop accurate methods estimate an individual's food energy intake. Measuring dietary intake open research problem. In particular, portion estimation challenging since process preparation consumption impose large variations on shapes appearances. this paper, we present method (kilocalories) from images using Generative Adversarial Networks (GAN). We introduce concept "energy distribution"...

This paper presents an integrated dietary assessment system based on food image analysis that uses mobile devices or smartphones. We describe two components of our system: a application and image-based nutrient database is connected to the application. An easy-to-use user interface described was designed preferences as well requirements methods. The validated by feedback collected from several studies. Food databases are also which facilitates enable dietitians other healthcare professionals...

Food image classification is challenging for real-world applications since existing methods require static datasets training and are not capable of learning from sequentially available new food images. Online continual aims to learn classes data stream by using each only once without forgetting the previously learned knowledge. However, none works target analysis, which more difficult incrementally due its high intra-class variation with unbalanced unpredictable characteristics future...

Advances in both lossy image compression and semantic content understanding have been greatly fueled by deep learning techniques, yet these two tasks developed separately for the past decades. In this work, we address problem of directly executing inference from quantized latent features compressed domain without pixel reconstruction. Although different methods proposed setting, they either are restrictive to a specific architecture, or sub-optimal terms task accuracy. contrast, propose...

The velocity of UAVs relative to the ground is usually measured using MEMS gyroscopes. In practice, output speed gyroscopes not only includes true UAVs, but also random white noise and outlier points. Therefore, how better process data by gyroscopes, remove outliers in data, great significance for improving accuracy providing good estimators[1]. order obtain more accurate information, this chapter first performs denoising processing on sensor based wavelet transform, analyzes measurement...

Idiopathic pulmonary fibrosis (IPF) is a progressive and irreversible lung disease with high mortality limited treatment options. While single-dose bleomycin-induced models are commonly used to investigate the pathogenesis of IPF, they fail adequately replicate complex pathological features in human patients, thereby hindering comprehensive investigations. Previous studies utilizing repetitive bleomycin injections have demonstrated closer resemblance IPF pathology; however, time-...