GoogLeNet-Based Ensemble FCNet Classifier for Focal Liver Lesion Diagnosis.

Transfer learning techniques are recently preferred for the computer aided diagnosis (CAD) of variety of diseases, as it makes the classification feasible from limited training dataset. In this work, an ensemble FCNet classifier is proposed to classify hepatic lesions from the deep features extracted using GoogleNet-LReLU transfer learning approachs. In the existing GoogLeNet architecture three modifications are done: ReLU activation functions in the inception modules are replaced by leaky ReLU activation function; a stack of three fully connected layers are included before the classification layer; and deep features of different level of abstraction extracted from the output of every inception layer given as classifier input in order to significantly enhance the classifier performance. The performance of the proposed classifier by the virtue of the above mentioned modifications is tested on six classes of liver CT images namely normal, hepatocellular carcinoma, hemangioma, cyst, abscess and liver metastasis. The results presented in this work demonstrate the efficacy of the proposed classifier design in achieving better classification accuracy.

1D Roughness Driven Depinning of Self-Assembly of Liquid Droplets.

The conventional hexagonal, uniform breath figure pattern formed over smooth substrates and substrates with constraints of the order of 50 µm is distorted when the underlying constraints are down to 1 µm. This paper explores this phenomenon further and concludes that, in addition to topology-based arguments presented by other authors previously, it is necessary to invoke the depinning effects of the three-phase contact line in order to explain the same. The influence of surface constraints on the self-assembly of liquid droplets is investigated. A semiquantitative explanation for large-scale pattern formation consisting of small-scale closely arranged droplets inside the large-scale distorted ring of droplets is presented in this paper. The scale at which the influence of constraints becomes dominant is also determined in this study. It is seen that the underlying roughness has a larger impact than the nature of polymer on pore size. Comparative studies of pore patterns formed on smooth and constrained substrates are reported. The simulated energy-minimized shapes of the droplets on smooth and constrained substrates are obtained using Surface Evolver.

A novel ketolide, RBx 14255, with activity against multidrug-resistant Streptococcus pneumoniae.

We present here the novel ketolide RBx 14255, a semisynthetic macrolide derivative obtained by the derivatization of clarithromycin, for its in vitro and in vivo activities against sensitive and macrolide-resistant Streptococcus pneumoniae. RBx 14255 showed excellent in vitro activity against macrolide-resistant S. pneumoniae, including an in-house-generated telithromycin-resistant strain (S. pneumoniae 3390 NDDR). RBx 14255 also showed potent protein synthesis inhibition against telithromycin-resistant S. pneumoniae 3390 NDDR. The binding affinity of RBx 14255 toward ribosomes was found to be more than that for other tested drugs. The in vivo efficacy of RBx 14255 was determined in murine pulmonary infection induced by intranasal inoculation of S. pneumoniae ATCC 6303 and systemic infection with S. pneumoniae 3390 NDDR strains. The 50% effective dose (ED50) of RBx 14255 against S. pneumoniae ATCC 6303 in a murine pulmonary infection model was 3.12 mg/kg of body weight. In addition, RBx 14255 resulted in 100% survival of mice with systemic infection caused by macrolide-resistant S. pneumoniae 3390 NDDR at 100 mg/kg four times daily (QID) and at 50 mg/kg QID. RBx 14255 showed favorable pharmacokinetic properties that were comparable to those of telithromycin.