Automatic segmentation of optic disc in retinal fundus images using semi-supervised deep learning

Diseases of the eye require manual segmentation and examination of the optic disc by ophthalmologists. Though, image segmentation using deep learning techniques is achieving remarkable results, it leverages on large-scale labeled datasets. But, in the field of medical imaging, it is challenging to acquire large labeled datasets. Hence, this article proposes a novel deep learning model to automatically segment the optic disc in retinal fundus images by using the concepts of semi-supervised learning and transfer learning. Initially, a convolutional autoencoder (CAE) is trained to automatically learn features from a large number of unlabeled fundus images available from the Kaggle’s diabetic retinopathy (DR) dataset. The autoencoder (AE) learns the features from the unlabeled images by reconstructing the input images and becomes a pre-trained network (model). After this, the pre-trained autoencoder network is converted into a segmentation network. Later, using transfer learning, the segmentation network is trained with retinal fundus images along with their corresponding optic disc ground truth images from the DRISHTI GS1 and RIM-ONE datasets. The trained segmentation network is then tested on retinal fundus images from the test set of DRISHTI GS1 and RIM-ONE datasets. The experimental results show that the proposed method performs on par with the state-of-the-art methods achieving a 0.967 and 0.902 dice score coefficient on the test set of the DRISHTI GS1 and RIM-ONE datasets respectively. The proposed method also shows that transfer learning and semi-supervised learning overcomes the barrier imposed by the large labeled dataset. The proposed segmentation model can be used in automatic retinal image processing systems for diagnosing diseases of the eye.

     

Investigating the Effect of WC on the Hardness and Wear Behaviour of Surface Modified AA 6063

The objective of this work is to reinforce Tungsten Carbide (WC) onto the surface of AA 6063 aluminium alloy using Gas Tungsten Arc (GTA) as heat source and investigate the hardness and wear properties. Based on number of trials, optimum GTA heat source parameters are finalized with reference to the proper fusion of base metal. It is found that the hardness is reduced after the application of heat. After the reinforcement of WC, the hardness returns back to the base metal value. In order to further to improve the properties, the reinforced alloy is heat Treated. The Microhardness of the reinforced and aged AA 6063 is increased by 50% with respect to base metal. The wear resistance of the reinforced surface is improved by 52% with respect to base metal. Characterization techniques like SEM, EDX and XRD are done and the presence of WC is confirmed.     

Fractionation of anhydrous milk fat by short-path distillation

Anhydrous milk fat was fractionated by short-path distillation into four fractions at temperatures of 245 and 265 C and pressures of 220 and 100 μm Hg. Two fractions (LF1 and LF2) were liquid, one fraction (IF) was semi-solid and one fraction (SF) was solid at room temperature. The fractions were characterized by melting temperature profile, solid fat index and triglyceride and fatty acid compositions. The peak melting temperature progressively increased (8.8 to 38.7 C) from liquid to solid fractions. The solid fat content ranged from 0 to 27.5% at 20 C, while native milk fat was 15.4%. The short chain (C24–C34) triglycerides were enriched in the LF1 fraction, long chain (C42–C54) triglycerides were concentrated in the SF fraction, and medium chain (C36–C40) triglycerides in the IF fraction; in the LF2 fraction, though, both short and medium chain triglycerides were enriched. Short chain (C4–C8) fatty acids gradually decreased from liquid to solid fractions and the trend was reverse for long chain (C14–C18) fatty acids, both saturated and unsaturated. The weight average molecular weights and geometric mean-carbon number of milk fat fractions were in the range of 590.7–782.8 and 31.9–46.3, respectively, compared to 729.3 and 41.0, respectively, for native milk fat, suggesting short-path distillation effects a very high degree of molecular weight separation.     

Self assembly of Co doped CeO2 microspheres from nanocubes by hydrothermal method and their photodegradation activity on AO7

We successfully synthesized self assembled Co doped CeO₂ microspheres from nanocubes via surfactant free hydrothermal technique. The surface shapes of CeO₂ nanoparticles have been controlled and optimized by adjusting the molar concentration and pH value of the precursor. The structural properties of the prepared nanocubes have been investigated using electron diffraction and X-ray diffraction. High resolution transmission electron microscopy studies show that the co doped CeO₂ nanocubes are in polycrystalline nature with highly porous structure. The photo catalytic activity of Co-doped CeO₂ nanocubes under UV–Visible light was examined by degradation of azodyes acid orange 7 (AO7) in aqueous solution. The improved photocatalytic activity was attributed to the better adsorption ability and unique 4f electron configuration of CeO₂.     

Mosquito-larvicidal efficacy of gold nanoparticles synthesized from the seaweed,Turbinaria ornata (Turner) J.Agardh 1848

Malaria is considered a dreadful mosquito-borne infectious disease of human beings caused and spread by biting of the female mosquito Anopheles stephensi infected with a parasitic protozoan Plasmodium falciparum. Continuous application of chemicals/synthetic insecticides for vector control causes various problems such as resistant mechanism of mosquito, toxicity to nontarget aquatic organisms and disturbance to the microbial community of the soil. Currently, green synthesized nanoparticles are being employed in various biological processes including insect and pest control. The present investigation focused on the mosquito-larvicidal property of Turbinaria ornata-mediated gold nanoparticles (To-AuNPs) and its boiled aqueous extract (To-AE) against the malarial vector A. stephensi. The recorded lethal concentration (LC₅₀ and LC₉₀) values (µg/ml) of To-AE and To-AuNPs against fourth instar larvae of A. stephensi were 37.77 and 159.55 and 12.79 and 78.70, respectively. The To-AuNPs were characterized through UV-visible spectroscopy, Fourier transform infrared spectroscopy (FTIR), x-ray diffraction (XRD), field emission scanning electron microscopy (FESEM), high-resolution transmission electron microscopy (HRTEM), energy dispersive x-ray spectroscopy (EDX), zeta potential and dynamic light scattering (DLS) method. The presently synthesized gold nanoparticles through the single-step, eco-friendly method is a potentially effective mosquitocidal agent.     

Hydrogenation of Vegetable Oils with Minimum trans and Saturated Fatty Acid Formation Over a New Generation of Pd-catalyst

Hydrogenation of sunflower and canola oils over a novel Pd-supported catalyst (pore size of 6.8 nm and BET specific surface area of 837 m²/g) was investigated and compared to commercial nickel catalyst. The formulated catalyst with Pd-loading of 1 wt%, supported on structured silica material was active and selective for the hydrogenation of sunflower and canola oils under mild process conditions. For both oils, the novel Pd supported catalyst exhibited a better selectivity than commercial Ni catalyst at a similar activity with a lower metal loading. For the same iodine value (IV) reduction, the Pd-catalyst produced less saturated fatty acids (SFA) and about the same level of trans fatty acids (TFA), but was more selective towards cis monoenes formation than Ni-catalyst. More importantly, this catalyst produced a reduced level of stearic acid, which at increased levels causes waxy mouth feel of the hydrogenated fat.