Deep learning facilitates complex medical data analysis and is increasingly being explored in colorectal cancer diagnostics. However, the training cost of the deep learning model limits its real-world medical utility. In this study, we present a composite network that combines deep learning and unsupervised K-means clustering algorithm (RK-net) for automatic processing of medical images. RK-net was more efficient in image refinement compared with manual screening and annotation. The training of a deep learning model for colorectal cancer diagnosis was accelerated by two times with utilization of RK-net-processed images. Better performance was observed in training loss and accuracy achievement as well. RK-net could be useful to refine medical images of the ever-expanding quantity and assist in subsequent construction of the artificial intelligence model.

With external optical feedback (EOF), a laser diode (LD) can operate at different dynamic states. In this work, an LD with EOF is set at period-one (P1) oscillation state to generate microwave photonic (MWP) signal for sensing. Firstly, the P1 state boundary of the LD is determined and then the influence of the LD controllable parameters on the boundary is studied by solving the well-known Lang-Kobayashi equations. A set of parameters selection rule for designing an LD based MWP sensing system is obtained. In addition, a measurement algorithm for recovering the displacement from an MWP sensing signal is developed. By making full use of the sensing information carried in both amplitude and frequency of the MWP signal, displacement sensing with high resolution and large range can be achieved. Both simulations and experiments are conducted to verify the proposed method and show it is capable of realizing wide measurable range, high measurement sensitivity, and high resolution sensing.

A novel Dual-frequency Doppler LiDAR (DFDL) is presented where the dual-frequency light source is generated by using external optical feedback (EOF) effect in a laser diode (LD). By operating a LD at period-one (P1) state and choosing suitable LD related parameters, a dual-frequency light source can be achieved. Such a dual-frequency source has advantages of the minimum part-count scheme, low cost in implementation, and ease in optical alignment. Theory and system design are presented for the proposed DFDL for velocity measurement with high measurement resolution. The proposed design has a potential contribution to the Light Detection And Ranging (LiDAR) in practical engineering applications.

In this study, a novel distance sensing method is presented by using a semiconductor laser (SL) with optical feedback (OF) and operating the SL at a switching status happened between two nonlinear dynamic states (stable state and period-one state). In this case, without the need for any electronic or optical modulation devices, the laser intensity can be modulated in a square wave form due to the switching via utilizing the inherent SL dynamics. The periodicity in the switching enables us to develop a new approach for long-distance sensing compared to other SL with OF-based distance measurement systems and lift the relevant restrictions that existed in the systems. Moreover, the impact of system controllable parameters on the duty cycle of the square wave signals generated was investigated on how to maintain the proposed system robustly operating at the switching status. Both simulation and experiment verified the proposed sensing approach.

Microwave photonic (MWP) sensing and measurement are envisioned to be a promising alternate to the conventional pure electronic or optical solutions. By using a semiconductor laser (SL) subjected to external optical feedback (EOF) operating at period-one (P1) dynamic state to generate MWP signals, a high performance sensing system is proposed for displacement with wide measurement range and super-high resolution. Through analysis on the state diagram of the system and its dynamics in P1, the relationship between the relaxation oscillation frequency of the system and external cavity length is determined. Further, a measurement formula for displacement sensing is worked out. Then, the relevant signal processing algorithm to retrieve displacement information is developed by considering mode-hopping, frequency-hopping, and sawtooth-like phenomena that occurred in the relationship. Both simulation and experiment results confirmed the proposed MWP sensing approach.

Waveform analysis is conducted on a sensing signal generated by a laser diode with dual-cavity operating at period-one state. The proposed design can achieve high measurement resolution and sensitivity for a moving object.