Food recognition using neural network classifier and multiple hypotheses image segmentation

ABSTRACT

This paper proposes the multiple-hypotheses image segmentation and feed-forward neural network classifier for food recognition to improve the performance. Initially, the food or meal image is given as input. Then, the segmentation is applied to identify the regions, where a particular food item is located using salient region detection, multi-scale segmentation, and fast rejection. Then, the features of every food item are extracted by the global feature and local feature extraction. After the features are obtained, the classification is performed for each segmented region using a feed-forward neural network model. Finally, the calorie value is computed with the aid of (i) food volume and (ii) calorie and nutrition measure based on mass value. The experimental results and performance evaluation are validated. The outcome of the proposed method attains 0.947 for Macro Average Accuracy (MAA) and 0.959 for Standard Accuracy (SA), which provides better classification performance.     

Heat transport in dilute mixtures of3He in superfluid4He

We report new measurements of the effective thermal conductivity K_eff and relaxation time τ in dilute mixtures of³He in superfluid⁴He, with molar concentrationsX≤10⁻³. The temperature range extended fromT≈1.4 K toT _λ. Both K_cff and τ are found to agree with theoretical predictions, in contrast to previous experiments where significant differences were observed. A new thermal conductivity cell design was used which almost completely eliminates extraneous volumes and surfaces, and the earlier results are explained in relation to these design changes.     

Development of Acoustic Instruments for Diagnosis and Management of Medical Conditions

Computerized analysis of lung sounds can play a very important role in management of patients with a variety of medical conditions, particularly those who are seriously ill. While single-channel devices can play an important role in making auscultation more objective, multi-channel instruments greatly improve the efficiency of data collection. This has a particular advantage in the rapid assessments of patients who are in acute distress from cardiopulmonary conditions. As noted, it has the distinct advantage of being noninvasive. This is a particular advantage in the case of children and pregnant women. A number of other groups are currently working with multichannel devices, but have not yet reported their findings. The future will likely see improved acoustic instruments useful in diagnosis and management of medical conditions     

Evaluation of Digitally Encoded Layer‐by‐layer Coated Microparticles as Cell Carriers

To obtain more biologically relevant data there is a growing interest in the use of living cells for assaying the biological activity of unknown chemical compounds. Density ‘multiplex’ cell‐based assays, where different cell types are mixed in one well and simultaneously investigated upon exposure to a certain compound are beginning to emerge. To be able to identify the cells they should be attached to microscopic carriers that are encoded. This paper investigates how digitally encoded microparticles can be loaded with cells while keeping the digital code in the microcarriers readable. It turns out that coating the surface of the encoded microcarriers with polyelectrolytes using the layer‐by‐layer (LbL) approach provides the microcarriers with a ‘highly functional’ surface. The polyelectrolyte layer allows the growth of the cells, allows the orientation of the cell loaded microcarriers in a magnetic field, and does not hamper the reading of the code. It has further been shown that the cells growing on the polyelectrolyte layer can become transduced by adenoviral particles hosted by the polyelectrolyte layer. It is concluded that the digitally encoded microparticles are promising materials for use in biomedical and pharmaceutical in‐vitro research where cells are used as tools.     

Chemokine receptors: structure, function and role in microbial pathogenesis

The chemokine superfamily is composed of at least 20 different leukocyte chemoattractants that act by binding to a family of G protein-coupled receptors. Leukocyte subtypes respond preferentially to unique but overlapping subsets of chemokines as determined by the receptor distribution, yet the receptors appear to signal through a common Gi-type G protein. Since chemokines appear to play major roles in inflammatory pathology, their receptors may be good targets for developing leukocyte selective anti-inflammatory drugs. Two chemokine receptors, CC CKRS and ONCC, function pathologically as cell entry factors respectively for human immunodeficiency virus 1, the cause of AIDS, and Plasmodium vivax, the major cause of malaria.