Browsing by Author "Ercan, B."

Filter results by typing the first few letters
Now showing 1 - 3 of 3
  • Thumbnail Image
    ItemUnknown
    Antibacterial properties and osteoblast interactions of microfluidically synthesized chitosan – SPION composite nanoparticles
    (Wiley Periodicals LLC, 2023-05-26) Kafali, M.; Şahinoğlu, O. Berkay; Tufan, Y.; Orsel, Z. C.; Aygun, Elif; Alyuz, Beril; Saritas, Emine Ulku; Erdem, E. Yegan; Ercan, B.
    In this research, a multi-step microfluidic reactor was used to fabricate chitosan – superparamagnetic iron oxide composite nanoparticles (Ch – SPIONs), where composite formation using chitosan was aimed to provide antibacterial property and nanoparticle stability for magnetic resonance imaging (MRI). Monodispersed Ch – SPIONs had an average particle size of 8.8 ± 1.2 nm with a magnetization value of 32.0 emu/g. Ch – SPIONs could be used as an MRI contrast agent by shortening T2 relaxation parameter of the surrounding environment, as measured on a 3 T MRI scanner. In addition, Ch – SPIONs with concentrations less than 1 g/L promoted bone cell (osteoblast) viability up to 7 days of culture in vitro in the presence of 0.4 T external static magnetic field. These nanoparticles were also tested against Staphylococcus aureus (S. aureus) and Pseudomonas aeruginosa (P. aeruginosa), which are dangerous pathogens that cause infection in tissues and biomedical devices. Upon interaction of Ch – SPIONs with S. aureus and P. aeruginosa at 0.01 g/L concentration, nearly a 2-fold reduction in the number of colonies was observed for both bacteria strains at 48 h of culture. Results cumulatively showed that Ch – SPIONs were potential candidates as a cytocompatible and antibacterial agent that can be targeted to biofilm and imaged using an MRI.
  • Thumbnail Image
    ItemUnknown
    High quality single-crystal germanium-on-insulator on bulk Si substrates based on multistep lateral over-growth with hydrogen annealing
    (American Institute of Physics, 2010-08-09) Yu, H. Y.; Cheng, S. L.; Park, J. H.; Okyay, Ali Kemal; Onbal, M. C.; Ercan, B.; Nishi, Y.; Saraswat, K. C.
    Germanium-on-insulator (GOI) is desired for high performance metal-oxide-semiconductor transistors and monolithically integrated optoelectronics. We demonstrate a promising approach to achieve single-crystal defect-free GOI by using lateral over-growth through SiO2 window. The dislocations due to the lattice mismatch are effectively terminated and reduced in SiO2 trench by selective area heteroepitaxy combined with hydrogen annealing. Low defect density of 4× 106 cm-2 and low surface roughness of 0.7 nm (root-mean-square) on GOI are confirmed by plan-view transmission electron microscopy and atomic force microscopy analysis. In addition, the excellent metal-semiconductor-metal diode electrical characteristics fabricated on this GOI confirm Ge crystal quality. The selectively grown GOI structure can provide the monolithic integration of SiGe based devices on a Si very large scale integration (VLSI) platform
  • Thumbnail Image
    ItemUnknown
    Synthetic18K: Learning better representations for person re-ID and attribute recognition from 1.4 million synthetic images
    (Elsevier, 2021-05-26) Aslan, C.; Ercan, B.; Ates, T.; Celikcan, U.; Erdem, A.; Erdem, E.; Üner, Onur Can
    Learning robust representations is critical for the success of person re-identification and attribute recognition systems. However, to achieve this, we must use a large dataset of diverse person images as well as annotations of identity labels and/or a set of different attributes. Apart from the obvious concerns about privacy issues, the manual annotation process is both time consuming and too costly. In this paper, we instead propose to use synthetic person images for addressing these difficulties. Specifically, we first introduce Synthetic18K, a large-scale dataset of over 1 million computer generated person images of 18K unique identities with relevant attributes. Moreover, we demonstrate that pretraining of simple deep architectures on Synthetic18K for person re-identification and attribute recognition and then fine-tuning on real data leads to significant improvements in prediction performances, giving results better than or comparable to state-of-the-art models.