Digital-based artificial neural network (ANN) machine learning is harnessed to reduce fiber nonlinearities, for the first time in ultra-spectrally-efficient optical fast orthogonal frequency division multiplexed (Fast-OFDM) signals. The proposed ANN design is of low computational load and is compared to the benchmark inverse Volterra-series transfer function (IVSTF)-based nonlinearity compensator. The two aforementioned schemes are compared for long-haul single-mode-fiber-based links at 9.69 Gb/s direct-detected optical Fast-OFDM signals. It is shown that an 80 km extension in transmission-reach is feasible when using ANN compared to IVSTF. This occurs because ANN can tackle stochastic nonlinear impairments, such as parametric noise amplification. Using ANN, the dynamic parameters requirements of the sub-ranging quantizers can also be relaxed compared to linear equalization, such as the reduction of the optimum clipping ratio and quantization bits by 2 dB and 2-bits, respectively, and by 2 dB and 2 bits when compared to the IVTSF equalizer.

Filter cascading effects are identified and compensated in flexible coherent multi‐carrier optical networks using, for the first time, a complex‐valued support vector machine (C‐SVM) non‐linear equaliser which is compared to phase‐conjugated subcarrier‐coding (PC‐SC)‐based non‐linearity compensation. The transmission performance of super‐channel‐based dual‐polarisation coherent multi‐band optical OFDM (DP‐MB‐OFDM) is analysed in a flexible network with two‐stage wavelength selective‐switch reconfigurable optical add‐drop multiplexer for coarse and fine switching granularity. It is shown that filter cascading effects have significant impact only on edge OFDM sub‐bands. On the other hand, C‐SVM outperforms PC‐SC resulting in an extension of the transmission‐reach, for instance, by 420 km when employing 32‐quadrature amplitude modulation at a targeted bit‐error‐rate of 10 ⁻³ . This occurs since C‐SVM can compensate more effectively stochastic‐induced nonlinear cross‐talk effects. Moreover, the interplay between polarisation‐mode dispersion (PMD) and fibre non‐linearity is also investigated in DP‐MB‐OFDM, revealing that a high PMD causes time fluctuation on the non‐linear interaction affecting the balance of the ‘twin‐subcarriers’ in PS‐SC.

Security issues of telemedicine‐based secure transmission of medical images find a very thin line drawn between diagnostic acceptability and cybersecurity. Partial but imperfect solutions emerge. JPEG2000 and HEVC concentrate only on compression, failing to provide any security consideration. Although secure, blockchain‐based systems introduce latency that impedes fine‐grained medical telepresence. The methods of homomorphic encryption are one of the very secure options, but they are almost impossible to carry out computationally. Watermarking schemes are usually incapable of providing real‐time detection of tampering. Given these drawbacks, such as missing real‐time tamper detection and poor integration between compression and security, as well as prohibitively high computational overhead, are our points to tackle; we recommend a hybrid framework based on adaptive compressed sensing (ACS), Secure Hash Algorithm 3 (SHA‐3), and lightweight encryption. The proposed framework obtains an improved CR of up to 30% better than JPEG2000 (13.5 bpp against 10.2 on x‐ray images), an 8.8% improvement in PSNR (43.2 vs. 39.7 dB), and a 6.6% increase in SSIM (0.97 compared to 0.91). Real‐time tampering detection (5.6 ms) safeguards the system against replay attacks (6.2 ms). When optimized for low‐latency transmission over constrained telemedicine scenarios, the algorithm shows greater efficiency and robustness than previously proposed methods.

This paper presents a novel integration of quadrature amplitude modulation (QAM)-based fast optical orthogonal frequency-division multiplexing (F-OFDM) with machine learning (ML)-based equalization in spatial division multiplexing (SDM) applications, using few-mode fibers (FMFs). The FMFs support four LP modes, resulting in a total of 12 orthogonal modes, each accommodating two polarizations. A digital multiple-input multiple-output channel equalizer is employed at the receiver’s digital signal processing (DSP) unit to effectively mitigate channel crosstalk. The study harnesses supervised ML-DSP techniques, in particular recurrent neural networks (RNNs) and deep neural networks (DNNs), achieving substantial reductions in bit error rates (BERs). In addition, higher-complexity architectures, namely convolutional neural networks (CNNs) and long short-term memory (LSTM) networks, are evaluated to assess the impact of advanced spatial and temporal feature extraction. It is shown that F-OFDM demonstrates superior performance over conventional optical OFDM, particularly when supported by ML techniques. Simulation results reveal that RNNs achieve a BER of 0.0019 over 15 km at 12 Gbaud (worst-case selected channel), showcasing a remarkable 52.5% improvement compared to linear equalization. DNNs achieve a BER of 0.0025, reflecting a 37.5% enhancement. While RNNs perform better, their computational demands pose challenges for real-time applications, and the more complex models (CNN and LSTM) do not provide additional performance gains. The paper also explores cyclic prefix management and subcarrier number strategies in F-OFDM to optimize performance, paving the way for future advancements in SDM networks.