Saliency-Aware Class-Agnostic Food Image Segmentation

Advances in image-based dietary assessment methods have allowed nutrition professionals and researchers to improve the accuracy of dietary assessment, where images of food consumed are captured using smartphones or wearable devices. These images are then analyzed using computer vision methods to estimate energy and nutrition content of the foods. Food image segmentation, which determines the regions in an image where foods are located, plays an important role in this process. Current methods are data dependent and thus cannot generalize well for different food types. To address this problem, we propose a class-agnostic food image segmentation method. Our method uses a pair of eating scene images, one before starting eating and one after eating is completed. Using information from both the before and after eating images, we can segment food images by finding the salient missing objects without any prior information about the food class. We model a paradigm of top-down saliency that guides the attention of the human visual system based on a task to find the salient missing objects in a pair of images. Our method is validated on food images collected from a dietary study that showed promising results.

Total Dietary Intake and Health Risks Associated with Exposure to Aflatoxin B1, Ochratoxin A and Fuminisins of Children in Lao Cai Province, Vietnam

The health burden of foodborne mycotoxins is considerable, but particularly for children due to their lower detoxification capacity, rapid growth and high intake of food in proportion to their weight. Through a Total Dietary Study approach, the objective was to estimate the dietary exposure and health risk caused by mycotoxins for children under 5 years living in the Lao Cai province in northern Vietnam. A total of 40 composite food samples representing 1008 individual food samples were processed and analyzed by ELISA for aflatoxin B1, ochratoxin A and fumonisins. Results showed that dietary exposure to aflatoxin B1, ochratoxin A and total fumonisins were 118.7 ng/kgbw/day, 52.6 ng/kg bw/day and 1250.0 ng/kg bw/day, respectively. Using a prevalence of hepatitis of 1%, the risk of liver cancer related to exposure of aflatoxin B1 was 12.1 cases/100,000 individual/year. Age-adjusted margin of exposure (MOE) of renal cancer associated with ochratoxin A was 127, while MOE of liver cancer associated with fumonisins was 542. Antropometric data show that 50.4% (60/119) of children were stunted, i.e. height/length for age z-scores (HAZ) below –2, and 3.4% (4/119) of children were classified as wasted, i.e. weight for height z-scores (WHZ) below –2. A significant negative relationship between dietary exposure to individual or mixture of mycotoxins and growth of children was observed indicating that the high mycotoxin intake contributed to stunning in the children studied.

Differences in the diet and trophic niche of three sympatric carnivores in a marshland

In order to conserve wetlands, it is crucial to have reliable knowledge of population and community processes. Our aims were to explore the dietary differences that allow coexistence of carnivores belonging to different trophic guilds, and to detect the presence of the relict Pannonian root vole subspecies (Microtus oeconomus méhelyi) within the Kis-Balaton marshland (Hungary). The 1-year dietary study was based on faecal analysis of the main mammalian predators of the area, i.e. red fox (Vulpes vulpes), martens (Martes spp.) and Eurasian otter (Lutra lutra). Small mammals, mainly Microtus species and water vole (Arvicola amphibius), were the primary foods for each carnivore. The dietary pattern of the otter was unique among European studies. Analysis of frequency data and estimated consumed biomass data revealed significant interspecific and intraspecific (seasonal) foraging differences among the three predators. Based on diets and live trapping of small mammals, foxes and martens selected voles, they selected against mice and shrews. The selection of the root vole was confirmed from faecal samples of foxes and martens. Consequently, faecal analysis of common carnivores can be an effective monitoring tool in providing data about prey species under special habitat conditions. Thus, it may contribute to wetland conservation indirectly.