Browsing by Subject "Triplet"
Now showing 1 - 4 of 4
Results Per Page
Sort Options
Item Open Access An analytical approach to the design of multiple mode rectangular cavity waveguide filters(Institute of Electrical and Electronics Engineers Inc., 2017) Kelleci, C.; Atalar, AbdullahThe multiple mode rectangular cavity structure with square corner cuts is revisited. An attempt to predict the physical dimensions of the cavity for dual mode second-order and triple mode third-order filters is made. Analytic expressions are formed to be used in the design process. The classical triple mode cavity filter structure is altered to give a finite frequency transmission zero either in the lower or upper sideband of the center frequency. The concept is illustrated with example designs. A novel additive manufacturing technique is used to fabricate a selected filter structure. The experimental results are in agreement with the expectations.Item Open Access Compact Ka-band filter applications based on the multiple mode rectangular cavity(Bilkent University, 2017-05) Kelleci, CeyhunFilters based on multiple mode cavity resonator technique have the advantage of realizing a given filter function in a reduced volume and weight with the drawback of increased complexity. In order to decrease the dependence on electromagnetic analysis software and to gain a better insight on the physics of the structure, the multiple mode single rectangular cavityfilter structure is investigated with an analytical approach. Expressions are obtained for the modal frequency shifts and for the intermodal coupling due to various types of corner cuts. An algorithm is proposed predicting the physical dimensions of thefinal structure given the corresponding coupling matrix. Example designs are realized. The algorithm is able to determine the physical dimensions of the second and third-orderfilters within a few percent. The classical triple mode rectangular cavityfilter structure is altered to form a triplet. The new triplet structure can be arranged to result in either a lower or higher sideband transmission zero. An example Ka-Band design is fabricated with both machining and a novel 3D printing technology. The results are in agreement with the expectations. Thefilter structure is further tailored to allow integration to Ka-Band waveguide output microwave modules without significant increase in the module's volume requirement.Item Open Access Content-based medical image retrieval with opponent class adaptive margin loss(Elsevier Inc., 2023-04-13) Öztürk, Şaban; Çelik, Emin; Çukur, TolgaThe increasing utilization of medical imaging technology with digital storage capabilities has facilitated the compilation of large-scale data repositories. Fast access to image samples with similar appearance to suspected cases in these repositories can help establish a consulting system for healthcare professionals, and improve diagnostic procedures while minimizing processing delays. However, manual querying of large repositories is labor intensive. Content-based image retrieval (CBIR) offers an automated solution based on quantitative assessment of image similarity based on image features in a latent space. Since conventional methods based on hand-crafted features typically show poor generalization performance, learning-based CBIR methods have received attention recently. A common framework in this domain involves classifier-guided models that are trained to detect different image classes. Similarity assessments are then performed on the features captured by the intermediate stages of the trained models. While classifier-guided methods are powerful in inter-class discrimination, they are suboptimally sensitive to within-class differences in image features. An alternative framework instead performs task-agnostic training to learn an embedding space that enforces the representational discriminability of images. Within this representational-learning framework, a powerful method is triplet-wise learning that addresses the deficiencies of point-wise and pair-wise learning in characterizing the similarity relationships between image classes. However, the traditional triplet loss enforces separation between only a subset of image samples within the triplet via a manually-set constant margin value, so it can lead to suboptimal segregation of opponent classes and limited generalization performance. To address these limitations, we introduce a triplet-learning method for automated querying of medical image repositories based on a novel Opponent Class Adaptive Margin (OCAM) loss. To maintain optimally discriminative representations, OCAM considers relationships among all image pairs within the triplet and utilizes an adaptive margin value that is automatically selected per dataset and during the course of training iterations. CBIR performance of OCAM is compared against state-of-the-art loss functions for representational learning on three public databases (gastrointestinal disease, skin lesion, lung disease). On average, OCAM shows an mAP performance of 86.30% in the KVASIR dataset, 70.30% in the ISIC 2019 dataset, and 85.57% in the X-RAY dataset. Comprehensive experiments in each application domain demonstrate the superior performance of OCAM against competing triplet-wise methods at 1.52%, classifier-guided methods at 2.29%, and non-triplet representational-learning methods at 4.56%.Item Open Access DNA codon recognition by a cubane wire: In silico approach(DergiPark, 2021) Mirzaei, M.; Hadipour, N.; Gülseren, OğuzDNA codons, consisting of triplet nucleotides (NTs), could play important roles for RNA transcription and protein translation in living systems. Therefore, their recognition could be seen important for diagnosis and therapy purposes. Based on triplet sequence formations of Adenine (A), Guanine (G), Cytosine (C) and Thymine (T) NTs, 64 codons were investigated in this work regarding their complexation with a molecular cubane (CUB) wire. To achieve this aim, each of singular 64 codons and CUB were optimized to be prepared for docking processes of complex formations. Hence, 64 complexes of codon-CUB were docked to see the recognition potency of CUB wire versus each of DNA codons. Interestingly, the obtained docking scores indicated that the CUB could work specifically versus the DNA codons, in which G-rich and A-rich triples were seen to be more favorable for complexation with CUB in comparison with other C-rich and T-rich triplet codons. Moreover, the results indicated that not pure G triplet but GAG codon was the most favorable one to be recognized by the CUB wire. However, pure T triplet was the worst one for such complex formations. The results of this work remarkably indicated that the CUB wire could work for recognition process of DNA codons from each other and such recognition could be very much specified for each of G-rich and A-rich codons, in which GAG codon was the best one among all the 64 investigated codons.