How Well Do Self-Supervised Models Transfer to Medical Imaging?

Self-supervised learning approaches have seen success transferring between similar medical imaging datasets, however there has been no large scale attempt to compare the transferability of self-supervised models against each other on medical images. In this study, we compare the generalisability of seven self-supervised models, two of which were trained in-domain, against supervised baselines across eight different medical datasets. We find that ImageNet pretrained self-supervised models are more generalisable than their supervised counterparts, scoring up to 10% better on medical classification tasks. The two in-domain pretrained models outperformed other models by over 20% on in-domain tasks, however they suffered significant loss of accuracy on all other tasks. Our investigation of the feature representations suggests that this trend may be due to the models learning to focus too heavily on specific areas.

Realizing UWB Antenna Array with Dual and Wide Rejection Bands Using Metamaterial and Electromagnetic Bandgaps Techniques.

This research article describes a technique for realizing wideband dual notched functionality in an ultra-wideband (UWB) antenna array based on metamaterial and electromagnetic bandgap (EBG) techniques. For comparison purposes, a reference antenna array was initially designed comprising hexagonal patches that are interconnected to each other. The array was fabricated on standard FR-4 substrate with thickness of 0.8 mm. The reference antenna exhibited an average gain of 1.5 dBi across 5.25-10.1 GHz. To improve the array's impedance bandwidth for application in UWB systems metamaterial (MTM) characteristics were applied it. This involved embedding hexagonal slots in patch and shorting the patch to the ground-plane with metallic via. This essentially transformed the antenna to a composite right/left-handed structure that behaved like series left-handed capacitance and shunt left-handed inductance. The proposed MTM antenna array now operated over a much wider frequency range (2-12 GHz) with average gain of 5 dBi. Notched band functionality was incorporated in the proposed array to eliminate unwanted interference signals from other wireless communications systems that coexist inside the UWB spectrum. This was achieved by introducing electromagnetic bandgap in the array by etching circular slots on the ground-plane that are aligned underneath each patch and interconnecting microstrip-line in the array. The proposed techniques had no effect on the dimensions of the antenna array (20 mm × 20 mm × 0.87 mm). The results presented confirm dual-band rejection at the wireless local area network (WLAN) band (5.15-5.825 GHz) and X-band satellite downlink communication band (7.10-7.76 GHz). Compared to other dual notched band designs previously published the footprint of the proposed technique is smaller and its rejection notches completely cover the bandwidth of interfering signals.

Improved adaptive impedance matching for RF front-end systems of wireless transceivers.

In this paper an automatic adaptive antenna impedance tuning algorithm is presented that is based on quantum inspired genetic optimization technique. The proposed automatic quantum genetic algorithm (AQGA) is used to find the optimum solution for a low-pass passive T-impedance matching LC-network inserted between an RF transceiver and its antenna. Results of the AQGA tuning method are presented for applications across 1.4 to 5 GHz (satellite services, LTE networks, radar systems, and WiFi bands). Compared to existing genetic algorithm-based tuning techniques the proposed algorithm converges much faster to provide a solution. At 1.4, 2.3, 3.4, 4.0, and 5.0 GHz bands the proposed AQGA is on average 75%, 49.2%, 64.9%, 54.7%, and 52.5% faster than conventional genetic algorithms, respectively. The results reveal the proposed AQGA is feasible for real-time application in RF-front-end systems.

Study on improvement of the performance parameters of a novel 0.41-0.47 THz on-chip antenna based on metasurface concept realized on 50 µm GaAs-layer.

A feasibility study is presented on the performance parameters of a novel on-chip antenna based on metasurface technology at terahertz band. The proposed metasurface on-chip antenna is constructed on an electrically thin high-permittivity gallium arsenide (GaAs) substrate layer. Metasurface is implemented by engraving slot-lines on an array of 11 × 11 circular patches fabricated on the top layer of the GaAs substrate and metallic via-holes implemented in the central patch of each row constituting the array, which connects the patch to the leaky-wave open-ended feeding slot-lines running underneath the patches. The slot-lines are connected to each other with a slit. A waveguide port is used to excite the array via slot-lines that couple the electromagnetic energy to the patches. The metasurface on-chip antenna is shown to exhibit an average measured gain in excess of 10 dBi and radiation efficiency above 60% over a wide frequency range from 0.41 to 0.47 THz, which is significant development over other on-chip antenna techniques reported to date. Dimensions of the antenna are 8.6 × 8.6 × 0.0503 mm3. The results show that the proposed GaAs-based metasurface on-chip antenna is viable for applications in terahertz integrated circuits.