Neuromorphic encryption decoding and brightness modulation display apparatus and method

Abstract

This invention belongs to the field of neuromorphic optoelectronics and liquid crystal optical readout technology. It proposes a neuromorphic encryption decoding and brightness modulation display device and method, comprising a photostimulation response unit, a signal conversion unit, a state generation unit, and a visual feedback unit. The photostimulation response unit uses photoelectric synaptic devices to generate a history-dependent electrical response under illumination by light pulses of different wavelengths. The signal conversion unit reads out and converts the electrical response to obtain a voltage signal or time-series signal that can be used for state characterization. The state generation unit discretizes the above signal into multi-level state codes. The visual feedback unit includes a polymer-dispersed liquid crystal display device and its driving circuit, which generates corresponding driving parameters based on the multi-level state codes to control the display device to output transmittance changes, grayscale patterns, or state prompts. This invention achieves a neuromorphic response closed loop from light input to visual feedback output.

Description

Technical Field

[0001] This invention relates to the field of neuromorphic optoelectronics and liquid crystal optical readout technology, and particularly to a neuromorphic encryption decoding and brightness modulation display device and method. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] Against the backdrop of the deep integration of information technology and display technology, neuromorphic optoelectronics provides an important pathway for building novel intelligent systems. Opto-synaptic devices can mimic the function of biological synapses, directly converting optical signals into historically dependent electrical responses and forming multi-level memory states through plasticity mechanisms, thereby achieving the integration of "sensing-storage-computing" at the physical level. Meanwhile, polymer-dispersed liquid crystals (PDLCs), as electro-optical modulation materials, can have their scattering or transmission states controlled by an external electric field, and are widely used in fields such as smart windows, privacy displays, and dynamic optical devices. Integrating opto-synapses with multispectral response capabilities with PDLC display units holds promise for constructing a closed-loop system where light input directly drives visual output, providing a new hardware paradigm for information security, human-computer interaction, and edge intelligence.

[0004] However, existing technologies still struggle to achieve a reliable mapping from multispectral optical stimulation to stable visual feedback within a single device. Specifically, when using the multi-level conductance states of multispectral photoelectric synapses to directly drive PDLCs for transmittance modulation or pattern display, there is a lack of an effective mechanism to discretize the historically dependent electrical response of the synaptic output into deterministic multi-level control signals that can be used to drive the PDLC. This deficiency makes it difficult to stably and repeatedly convert the analog output of the synapse into the driving parameters required by the PDLC, hindering the formation of a reliable neuromorphic response closed loop from optical input to visual pattern output. Consequently, this limits the practical application of such integrated devices in scenarios such as encryption decoding and dynamic display. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a neuromorphic encryption decoding and brightness modulation display device and method, which realizes full-chain control of multispectral input, state formation, quantization encoding, decoding decision and visual output, reduces external links, lowers the risk of data leakage, and reduces power consumption and response latency compared to traditional discrete architectures.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a neuromorphic encryption decoding and brightness modulation display device.

[0007] A neuromorphic encryption decoding and brightness modulation display device includes a light stimulus response unit, a signal conversion unit, a state generation unit, and a visual feedback unit; The photostimulation response unit includes photosynaptic devices for generating a history-dependent electrical response under illumination by light pulses of different wavelengths; The signal conversion unit is electrically connected to the photostimulation response unit and is used to read out and convert the electrical response to obtain a voltage signal or time series signal that can be used for state characterization. The state generation unit is connected to the signal conversion unit and is used to discretize voltage signals or time series signals into multi-level state codes. The visual feedback unit is signal-connected to the state generation unit. The visual feedback unit includes a PDLC display device and a driving circuit. It is used to generate corresponding driving parameters according to multi-level status codes and control the output transmittance change, grayscale pattern and status prompt of the PDLC display device to realize the neuromorphic response closed loop from light input to visual feedback output.

[0008] In one implementation of the first aspect of the present invention, the photoelectric synaptic device adopts an HfAlO / NbOx heterostructure, in which the HfAlO trap control layer and the NbOx photosensitive layer form a heterostructure interface. The area near the heterostructure interface contains oxygen vacancy-related defect trap states and interface trap states to achieve the capture and release of photogenerated carriers, thereby generating short-term or long-term plasticity.

[0009] In one implementation of the first aspect of the present invention, the photoelectric synaptic device generates a wavelength-selective electrical response to light stimulation in the 250-500 nm band, and forms more than 16 discrete conductive states under excitation at least two wavelengths among 350 nm, 450 nm and 500 nm. The state generation unit generates multi-level state codes based on at least one of the following: peak value, steady-state value, integral value, decay time constant, and pulse sequence statistical characteristics.

[0010] In one implementation of the first aspect of the present invention, the preset optical key conditions include at least one of the following: wavelength combination, pulse width, number of pulses, pulse frequency, pulse interval, and wavelength application order. The photostimulation response unit includes a first synaptic channel and a second synaptic channel in parallel. The first synaptic channel is used to encode color information, and the second synaptic channel is used to encode structural information, brightness information, or watermark information, forming a hierarchical encoding or multi-factor key encoding. When the preset optical key conditions are met, a valid decoding result is output; when the preset optical key conditions are not met, a rejection state is output.

[0011] In one implementation of the first aspect of the present invention, the signal conversion unit includes a transimpedance amplifier circuit and a charge amplifier circuit, and includes a baseline correction unit, a normalization unit and a noise suppression unit. The state generation unit includes a discretization mapping unit, which is used to discretize the electrical response into a multi-level rating code and map it into PDLC driving parameters to form a token interface that can directly drive the display. The visual feedback unit also includes a decoding module, which is signal-connected to the state generation unit. The decoding module includes a reservoir calculation unit and a classification decision unit for outputting the decoding results.

[0012] In one implementation of the first aspect of the present invention, the PDLC display device includes a PDLC thin film layer sandwiched between two transparent ITO electrodes, and the driving signal output by the driving circuit is an AC driving signal or a pulse driving signal, so that the PDLC display device has a threshold range and a saturation range, and realizes continuous brightness modulation or graded grayscale output. The ITO electrode is a patterned electrode or a partitioned electrode; the PDLC thin film layer is a liquid crystal microdroplet-polymer network structure, and the polymerizable monomer system includes at least one of a crosslinking agent, an acrylic monomer, and doped nanoparticles.

[0013] Secondly, the present invention provides a neuromorphic encryption decoding and brightness modulation display method.

[0014] A neuromorphic encryption decoding and brightness modulation display method, utilizing the neuromorphic encryption decoding and brightness modulation display device of the first aspect of the present invention, includes the following process: Set preset optical key conditions and map the information to be processed into a multi-wavelength optical pulse sequence; Apply a sequence of light pulses and obtain the electrical response; The electrical response is read out, conditioned, and quantized into a grade code or feature vector to generate an encoded result. When the preset optical key conditions are met, the encoding result is decoded to obtain the decoding result; The encoded or decoded result is mapped to a PDLC drive signal and loaded onto the PDLC display device to output a brightness pattern. When the preset optical key conditions are not met, a rejection status is output and the PDLC display device is controlled to output a rejection pattern or a blank pattern.

[0015] In one implementation of the second aspect of the present invention, decoding includes calculating a state vector through a reservoir and classifying or reconstructing the state vector through a classification decision device; the preset optical key conditions include at least a joint constraint on the wavelength combination and the order of wavelengths. The PDLC drive signal is obtained by mapping a level code or state vector, and the mapping parameters include at least the drive voltage amplitude level; brightness enhancement, brightness compression and graded grayscale output are achieved by configuring the drive operating point near the threshold range or near the saturation range.

[0016] Thirdly, the present invention provides a method for preparing a neuromorphic encryption decoding and brightness modulation display device.

[0017] A method for fabricating a neuromorphic encryption decoding and brightness modulation display device includes the following steps: A substrate is provided and cleaned to form a first electrode and a second electrode; Formation of HfAlO trap control layer; A NbOx photosensitive layer is formed to construct an HfAlO / NbOx heterostructure, and interface trap states or defect trap states are introduced near the heterostructure by annealing or oxygen partial pressure regulation during deposition. Encapsulate multispectral opto-synaptic units and connect them to various functional modules; A PDLC thin film layer was fabricated and sandwiched between two transparent ITO electrodes to form a PDLC brightness modulation display module; Establish the mapping relationship between the output of the multispectral photoelectric synapse unit and the PDLC driving signal and complete the integration.

[0018] In one implementation of the third aspect of the present invention, a PDLC thin film layer is prepared and sandwiched between two transparent ITO electrodes to form a PDLC brightness modulation display module, comprising: A liquid crystal phase is mixed with a polymerizable monomer system and a photoinitiator is added. The mixture is then coated or infused, gap-controlled and UV-cured to form a PDLC thin film layer with a liquid crystal microdroplet-polymer network structure. The polymerizable monomer system includes at least one of a crosslinking agent, an acrylic monomer or doped nanoparticles. Establish and integrate the mapping relationship between the multispectral photoelectric synapse unit output and the PDLC drive signal, including: A one-to-one or segmented correspondence is established between the multi-level discrete states of the multispectral photoelectric synapse unit and the PDLC driving parameters. Mapping parameters or mapping tables are established for different wavelength channels to output a target brightness pattern when the preset optical key conditions are met, and a rejection pattern or blank pattern is output when the conditions are not met.

[0019] Compared with the prior art, the beneficial effects of the present invention are: This invention innovatively proposes a neuromorphic encryption decoding and brightness modulation display device, effectively solving the technical problem in existing technologies where the analog output of photosynapses cannot be stably and reproducibly converted into liquid crystal display driving parameters. This solution introduces a signal conversion unit and a state generation unit to accurately read out, convert, and discretize the historically dependent electrical response generated by the photostimulation response unit into multi-level state codes. This series of processing steps constructs a standardized "token" interface, transforming the originally continuous and easily interfered synaptic electrical signals into clear, discrete digital instructions. The PDLC display device and its driving circuit in the visual feedback unit directly receive these multi-level state codes and generate precisely corresponding driving parameters, thereby reliably controlling the preset transmittance changes, grayscale patterns, or status prompts of the PDLC output. Thus, a neuromorphic response closed loop from multispectral light input to stable visual feedback output is successfully established within a single device, ensuring the reliability, reproducibility, and practicality of the entire system.

[0020] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0021] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0022] Figure 1 A schematic diagram illustrating the principle of multi-spectral photosynapse outputting multi-level current and driving PDLC to achieve brightness modulation display, provided as an exemplary embodiment of the present invention; Figure 2 X-ray photoelectron spectroscopy (XPS) characterization spectrum of the HfAlO trap control layer provided as an exemplary embodiment of the present invention; Figure 3 A schematic diagram illustrating the analogy between biological synapses and multispectral photoelectric synapse devices, and a schematic diagram of the layered structure of an HfAlO / NbOx photoelectric synapse device, provided as an exemplary embodiment of the present invention; Figure 4 A schematic diagram illustrating the state evolution of a photosynaptic unit under multispectral light pulse stimulation, as provided in an exemplary embodiment of the present invention; Figure 5 Microstructure characterization diagram of group A PDLC samples provided as an exemplary embodiment of the present invention; Figure 6 Microstructure characterization diagram of group B PDLC samples provided as an exemplary embodiment of the present invention; Figure 7 Microstructure characterization diagram of group C PDLC samples provided as an exemplary embodiment of the present invention; Figure 8 A schematic diagram of the excitatory postsynaptic current (EPSC) response curve under a single light pulse stimulus, provided as an exemplary embodiment of the present invention; Figure 9 A schematic diagram of the paired pulse facilitated (PPF) response curve under dual-pulse stimulation provided as an exemplary embodiment of the present invention; Figure 10 A schematic diagram of the accumulation and decay response curves of excitatory postsynaptic currents (EPSC) under different pulse width conditions provided as an exemplary embodiment of the present invention; Figure 11 A schematic diagram of the accumulation and retention response curves of excitatory postsynaptic currents (EPSC) under different pulse number conditions provided as an exemplary embodiment of the present invention; Figure 12 A schematic diagram of excitatory postsynaptic current (EPSC) response curves under different pulse frequencies provided as an exemplary embodiment of the present invention; Figure 13 A schematic diagram of a multi-level discrete current state distribution formed under different wavelength excitation conditions, provided as an exemplary embodiment of the present invention; Figure 14 A schematic diagram of the system architecture and signal flow of the apparatus provided in an exemplary embodiment of the present invention; Figure 15 A comparison diagram of the electro-optical response curve of a PDLC brightness modulation display module and its display effect under different driving voltages, provided as an exemplary embodiment of the present invention; Figure 16 An example flowchart and a comparison diagram of recognition results of the apparatus provided in an image processing application, which is an exemplary embodiment of the present invention. Detailed Implementation

[0023] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0024] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0025] This embodiment provides a neuromorphic encryption decoding and brightness modulation display device. For example... Figure 1As shown, this device completes closed-loop processing at the edge, including multi-wavelength light input, synaptic state formation, grade code generation, decoding decision, and PDLC brightness output, reducing external data handling and intermediate link risks. The device includes a photostimulation response unit, a signal conversion unit, a state generation unit, and a visual feedback unit. These units form a processing link through electrical and signal connections. The photostimulation response unit includes HfAlO / NbO. x The photoelectric synaptic device generates a history-dependent current response under illumination by light pulses of different wavelengths. A signal conversion unit, electrically connected to the photostimulation response unit, includes a transimpedance amplifier circuit to achieve current-to-voltage conversion and includes a filtering unit, a baseline correction unit, and a normalization unit to output a quantifiable voltage signal. A state generation unit, signal-connected to the signal conversion unit, includes a feature extraction unit and a discretization mapping unit to discretize the voltage signal into multi-level state codes. A visual feedback unit, signal-connected to the state generation unit, includes a PDLC display device and a driving circuit, generating driving parameters based on the state codes and outputting transmittance changes, grayscale patterns, or status prompts. A decoding and decision module, signal-connected to the state generation unit, includes a reservoir calculation unit and a classification decision unit for decoding decisions and key verification.

[0026] like Figure 3 As shown, the HfAlO / NbOx photoelectric synapse device adopts a layered heterostructure, consisting of the following layers from bottom to top: a glass substrate, a Pt bottom electrode (50 nm thick), an HfAlO trapping and modulation layer (10 nm thick), an NbOx photosensitive layer (20 nm thick), and an ITO transparent top electrode (50 nm thick). The NbOx photosensitive layer absorbs multi-wavelength photons to generate photogenerated carriers, while the HfAlO trapping and modulation layer provides trapped states at the heterostructure interface, enabling carrier capture and release, and generating a history-dependent electrical response. Figure 3 As shown in the right figure, this structure is analogous to the presynaptic membrane, synaptic cleft, and postsynaptic membrane of a biological synapse: the light pulse corresponds to the presynaptic neuron signal, the HfAlO / NbOx heterogeneous interface corresponds to the neurotransmitter transmission and receptor response in the synaptic cleft, and the current output corresponds to the postsynaptic current.

[0027] The fabrication process of HfAlO / NbOx photosynaptic devices includes: Step A1: Substrate treatment and preparation of the first electrode.

[0028] A glass substrate (25mm × 25mm, 1mm thickness) was selected and ultrasonically cleaned sequentially with acetone, isopropanol, and deionized water for 10 min each, followed by nitrogen drying. A Pt thin film was deposited on the substrate as the first electrode using physical vapor deposition (PVD). Deposition parameters: background vacuum 5 × 10⁻⁶. -4A Pt bottom electrode with a thickness of 50 nm was obtained by sputtering at Pa, Ar gas flow rate of 20 sccm, sputtering power of 100 W, and deposition time of 400 s. The effective area was defined as 5 mm × 5 mm by using a mask.

[0029] Step A2: Preparation of HfAlO trap control layer.

[0030] HfAlO films were deposited on Pt substrates using plasma-enhanced atomic layer deposition (ALD-plasma). Deposition parameters were: substrate temperature 250 °C, Hf precursor Hf[N(CH3)(C2H5)]4, Al precursor Al(CH3)3, oxygen plasma power 300 W, and pulse duration 0.1 s. Alternating Hf and Al deposition cycles of 25 cycles were employed, with a growth rate of approximately 0.1 nm / cycle and a total thickness of 10 nm. The composition and chemical states of the HfAlO films were characterized by X-ray photoelectron spectroscopy (XPS), such as... Figure 2 As shown: The full spectrum shows the presence of Hf, Al, and O elements, while the Hf 4f fine spectrum shows Hf 4+ Al 2p fine spectrum shows Al (binding energy approximately 17 eV) in the state of Al 2p (binding energy approximately 17 eV). 3+ The O 1s fine spectrum fitting showed an oxygen vacancy-related peak (binding energy approximately 74 eV) and a metal-oxygen bond peak (binding energy approximately 531 eV), verifying the presence of oxygen vacancy-related defects in the HfAlO film.

[0031] Step A3: Preparation of NbOx photosensitive layer and manipulation of trap states.

[0032] NbOx films were deposited on HfAlO films using PVD. Deposition parameters: background vacuum 5 × 10⁻⁶. -4 A 20 nm thick NbOx photosensitive layer was obtained by sputtering at 15 sccm of Ar gas and 5 sccm of O2 gas (oxygen partial pressure approximately 25%), with a sputtering power of 80 W and a deposition time of 1250 s. After deposition, an annealing process was performed in a tube furnace to control the interface trap states: annealing temperature 300 °C, annealing time 30 minutes, nitrogen atmosphere, flow rate 100 sccm. The annealing process promoted the redistribution of oxygen vacancies near the interface, forming interface trap states and defect trap states, enhancing the photogenerated carrier trapping capability.

[0033] Step A4: Preparation and packaging of the second electrode.

[0034] An ITO transparent electrode was deposited on a NbOx thin film using PVD as the second electrode. Deposition parameters: Base vacuum 5 × 10⁻⁶ -4A 50 nm thick ITO top electrode was obtained by sputtering at 100 W with an Ar gas flow rate of 20 sccm, an O2 gas flow rate of 2 sccm, a sputtering power of 100 W, and a deposition time of 500 s. The effective area of ​​the ITO electrode was limited to 4 mm × 4 mm using a mask to ensure a safe light incident window. Silver paste was used to create leads on the Pt bottom electrode and the ITO top electrode, and the device edges were encapsulated with epoxy resin to prevent oxidation and moisture intrusion.

[0035] After packaging, the device underwent electrical shaping: a ±1 V scan voltage was applied 10 times in the dark to stabilize the initial conductivity state. Subsequently, optical calibration was performed: pulses with a width of 500 ms and an intensity of 1 mW / cm² were applied at wavelengths of 350 nm, 450 nm, and 500 nm, respectively. 2 The standard optical pulse is used to measure the peak value, steady-state value and decay time constant of the excitatory postsynaptic current (EPSC), and to establish a "wavelength-pulse parameter-output state" lookup table for subsequent key constraints and decoding decisions.

[0036] The working mechanism of HfAlO / NbOx photosynaptic devices specifically includes: like Figure 4 As shown, under multi-wavelength light pulse stimulation, the NbOx photosensitive layer absorbs photons to generate photogenerated electron-hole pairs, and the photogenerated carriers migrate to the HfAlO / NbOx heterostructure interface. Oxygen vacancy-related defect trap states and interface trap states near the interface trap the carriers, modulating the local barrier height and causing a cumulative enhancement in the current response (corresponding to short-term plasticity). Under continuous light stimulation, the trap states gradually saturate, and the current reaches a steady state (corresponding to long-term plasticity). After the light is removed, the trapped carriers are slowly released, and the current decays exponentially; the decay time constant reflects the depth distribution of the trap states. By controlling the HfAlO thickness, annealing temperature, and oxygen partial pressure, the trap density and energy level distribution can be adjusted to achieve different responses at different wavelengths, expanding the encoding dimension.

[0037] The fabrication process of the PDLC display module specifically includes: Two glass substrates (30 mm × 30 mm, 1.1 mm thick) were selected, and ITO transparent electrodes with a thickness of 150 nm were deposited on the substrates by magnetron sputtering. A patterned ITO electrode was fabricated on one of the substrates using photolithography and wet etching processes. The pattern was the character "Correct" with a line width of 2 mm, used to display area markings.

[0038] Polyimide (PI) alignment layers, approximately 50 nm thick, were coated onto the ITO electrode surfaces of two substrates and unidirectionally aligned through a triboelectric treatment. Spherical spacers, approximately 10 μm in diameter and at a density of approximately 100 spacers / cm², were uniformly distributed on one of the substrates. 2Two substrates are bonded together with ITO electrodes facing each other, and the cell thickness is controlled to be 10 μm by using spacers.

[0039] The PDLC precursor solution was prepared with the following formulation: liquid crystal phase E7 (Merck) 70 wt%, crosslinking agent UV6300 (Nippon Kayaku Co., Ltd.) 15 wt%, acrylic monomer THFMA (Sigma-Aldrich) 10 wt%, AZO nanoparticles (average particle size 20 nm) 3 wt%, and photoinitiator Irgacure 651 (BASF) 2 wt%. The precursor solution was stirred at room temperature for 2 hours to ensure uniform mixing.

[0040] The precursor solution was injected into the gap between the two substrates using a vacuum infusion method at a pressure of approximately 0.1 Pa for 10 minutes. After infusion, the device was cured under a 365 nm ultraviolet light source with an intensity of 10 mW / cm². 2 The curing time is 8 minutes. During the curing process, the polymerizable monomers undergo photopolymerization to form a polymer network, and the liquid crystal phase is dispersed into microdroplets, forming a liquid crystal microdroplet-polymer network structure.

[0041] After curing, the edges of the device are sealed with epoxy resin to prevent liquid crystal leakage. Wires are then led out from the ITO electrodes to complete the fabrication of the PDLC display device.

[0042] The electro-optical performance characterization of the PDLC display module specifically includes: like Figure 5 , Figure 6 and Figure 7 As shown, the microstructure and electro-optic properties of PDLC samples with different formulations (groups A, B, and C) were characterized. Group A samples contained only the liquid crystal phase E7 and the crosslinking agent UV6300. Scanning electron microscopy (SEM) morphology images showed that the liquid crystal droplets were approximately 5-10 μm in size and unevenly distributed. The transmittance-voltage curves showed that the transmittance was approximately 10% at 0 V (low transmittance / high scattering state), and the threshold voltage V... th Approximately 9 V, saturation voltage V sat Approximately 27 V, saturation transmittance approximately 70%, contrast ratio (saturation transmittance / 0 V transmittance) approximately 7. Response time: rise time (10%-90% transmittance) approximately 65 ms, fall time (90%-10% transmittance) approximately 70 ms.

[0043] Group B samples, with the addition of THFMA to Group A, showed that the liquid crystal microdroplet size was reduced to 3-6 μm and the distribution was more uniform under scanning electron microscopy (SEM). Transmittance-voltage curves showed the threshold voltage V... th Approximately 11 V, saturation voltage V satApproximately 32V, contrast ratio approximately 10. Response time: rise time approximately 50 ms, fall time approximately 55 ms. Adding acrylic monomers increases crosslinking density, reduces liquid crystal droplet size, raises threshold voltage, and improves contrast and response speed.

[0044] Sample C, with the addition of AZO nanoparticles to the sample from group B, showed a further reduction in liquid crystal droplet size to 2-4 μm and the most uniform distribution, as revealed by scanning electron microscopy (SEM). Transmittance-voltage curves showed a threshold voltage (Vth) of approximately 13 V, a saturation voltage (Vsat) of approximately 37 V, and a contrast ratio of approximately 12. Response time was approximately 40 ms rise time and 45 ms fall time. The addition of nanoparticles increased the dielectric constant, further raised the threshold voltage, and significantly improved the contrast ratio and response speed.

[0045] In summary, the formulation in group C (containing crosslinking agent + acrylic monomer + nanoparticles) has the highest threshold voltage, the highest contrast ratio, and the fastest response time, and is therefore the preferred option in this embodiment.

[0046] In this implementation, a mapping relationship needs to be established between the multispectral photosynapse output and the PDLC drive signal. The signal conversion unit includes a transimpedance amplifier circuit, which is composed of an operational amplifier (such as OPA128) and a feedback resistor (1 MΩ), with a gain of 10. 6 The V / A converter transforms the nA-μA level current output from the photosynaptic device into a mV-V level voltage. The output of the transimpedance amplifier circuit is filtered by filter unit 22 (a second-order low-pass filter with a cutoff frequency of 1kHz) to remove high-frequency noise. It then undergoes processing by baseline correction unit 23 (subtracting the dark-state baseline voltage) and normalization unit 24 (dividing by the peak voltage) to obtain the normalized voltage signal. .

[0047] The state generation unit includes a feature extraction unit 1 and a discretization mapping unit 2. Feature extraction unit 1... Peak Extraction steady-state value Integral value decay time constant (Obtained through exponential fitting) and other feature quantities. Discretization mapping unit 2 maps the feature quantities to 4-bit level codes (16 levels, numbered 0-15) according to the calibration threshold. Calibration process: Apply standard light pulses with a pulse width of 500 ms and a light intensity of 1 mW / cm² 20 times each at wavelengths of 350 nm, 450 nm, and 500 nm, and collect data. After normalization, the values ​​are sorted from smallest to largest and divided into 16 threshold intervals using equal intervals: [0, 0.0625], [0.0625, 0.125], ..., [0.9375, 1]. This establishes a threshold range... -Level Code Mapping Table.

[0048] The decoding and decision module is signal-connected to the state generation unit, including the reservoir computing unit and the classification decision unit. The reservoir computing unit adopts an echo state network (ESN) structure, containing 300 reservoir nodes, and the input weight matrix... (300×1), Reserve Pool Weight Matrix (300×300, spectral radius 1.0), output weight matrix (1×300). The input is the grade code sequence {k(t)}, and the reservoir state update equation is: Output state vector The classification decision unit uses a linear support vector machine (SVM) to perform binary classification (target / rejection) on y(t). The decision threshold is determined through the training set. When the preset optical key conditions are met (e.g., wavelength sequence of 350 nm-450 nm-500 nm, pulse interval of 500 ms, and 5 pulses of each), y(t) > the threshold, and a valid decoding result is output; when the conditions are not met, y(t) < the threshold, and a rejection state is output.

[0049] Establish a segmented mapping table for "level code - driving voltage". The transmittance-voltage curve of the PDLC display device (e.g., ...) Figure 15 (As shown on the left) has a threshold range (V) th =13 V to V th +5 V=18 V) and saturation range (V sat =37 V to V sat +5 V = 42V). Segmentation mapping rule: Low-level codes (levels 0-7) are mapped to the threshold range, driving voltage. V and k are the level codes; higher level codes (levels 8-15) are mapped to the saturation interval. V. The driving signal is a square wave pulse with a frequency of 1 kHz, a duty cycle of 50%, and an amplitude of [missing value]. For different wavelength channels, a mapping table is established so that the target brightness pattern (such as the character "Correct") is output when the key conditions are met, and a blank pattern (0V driven) is output when the conditions are not met.

[0050] In this implementation, synaptic plasticity is characterized, specifically including: After completing device fabrication and system integration, the photoelectric synaptic device was characterized for synaptic plasticity, and a set of quantization thresholds and key parameters was established. Test circuit: The Pt bottom electrode was grounded, and the ITO top electrode was connected to an oscilloscope via a transimpedance amplifier circuit to acquire excitatory postsynaptic current (EPSC) time curves. Optical stimulation: An LED light source (wavelength switchable at 350 nm, 450 nm, and 500 nm, light intensity 1 mW / cm²) was used. 2The pulse width, interval, number, and frequency are controlled by a function generator.

[0051] like Figure 8 As shown, under the action of a single light pulse with a pulse width of 500 ms and a light intensity of 1 mW / cm², the excitatory postsynaptic current (EPSC) rapidly rises to a peak value A1 (approximately 120 nA), and then slowly decays to a steady-state value (approximately 20 nA). The peak value A1 reflects the initial photogenerated carrier concentration and trapping rate, serving as a fundamental feature for subsequent quantization encoding.

[0052] like Figure 9 As shown, under the action of two optical pulses with a pulse width of 500 ms and an interval of Δt = 500 ms, the peak value of the second response A2 (approximately 150 nA) is higher than that of A1, and the Paired Pulse Facilitation (PPF) ratio A2 / A1 is approximately 1.25. The Paired Pulse Facilitation (PPF) ratio increases as Δt decreases (A2 / A1 is approximately 1.5 when Δt = 100 ms), reflecting a short-term memory timescale of approximately 1 s. The Paired Pulse Facilitation (PPF) feature can be used as a time-dimensional key parameter; an incorrect interval cannot trigger the correct Paired Pulse Facilitation (PPF) ratio, thus reducing the false trigger rate.

[0053] The effect of pulse width on excitatory postsynaptic currents (EPSC), such as Figure 10 As shown, specifically, pulse widths (100 ms, 500 ms, 1000 ms) were varied, and a pulse sequence (5 pulses, spaced 500 ms apart) was applied. With a pulse width of 100 ms, the cumulative enhancement of excitatory postsynaptic current (EPSC) was relatively weak, with a peak value of approximately 80 nA and a relatively rapid decay; with a pulse width of 500 ms, the peak value was approximately 120 nA, and the accumulation was significant; with a pulse width of 1000 ms, the peak value was approximately 180 nA, the accumulation was strongest, and the decay was slowest. The pulse width can adjust the equivalent memory strength, serving as a key parameter or encoding dimension.

[0054] The effect of pulse number on excitatory postsynaptic currents (EPSC), such as Figure 11 As shown, specifically, the number of pulses (1, 5, 10, 20), pulse width 500 ms, and interval 500 ms were varied. With 1 pulse, the peak value was approximately 120 nA, decaying to baseline in about 2 seconds; with 5 pulses, the peak value was approximately 180 nA, decaying slowly to about 5 seconds; with 10 pulses, the peak value was approximately 220 nA, decaying in about 10 seconds; and with 20 pulses, the peak value was approximately 250 nA, decaying in about 20 seconds. Increasing the number of pulses enhanced the accumulation of excitatory postsynaptic current (EPSC) and improved retention capacity, which can be used as a basis for encoding dimension and threshold calibration.

[0055] The effect of pulse frequency on excitatory postsynaptic currents (EPSC), specifically, such as... Figure 12 As shown, the pulse frequency (1 Hz, 10 Hz, 100 Hz), pulse width (100 ms), and number of pulses were varied. At 1 Hz, the peak value was approximately 100 nA, with relatively weak accumulation; at 10 Hz, the peak value was approximately 150 nA, with significant accumulation; and at 100 Hz, the peak value was approximately 200 nA, with the strongest accumulation and slowest decay. At higher frequencies, the pulse intervals were shorter, and the trapped states were recaptured before being fully released, resulting in enhanced accumulation. The pulse frequency can be used as a time-dimensional key parameter to adjust the state spacing.

[0056] In summary, the effects of single-pulse response, paired-pulse facilitation (PPF), pulse width, number, and frequency on excitatory postsynaptic current (EPSC) are clearly defined, providing a basis for subsequent quantization thresholds, grade code mapping, and key parameter sets.

[0057] like Figure 13 As shown, at wavelengths of 350 nm, 450 nm, and 500 nm, by varying the number of pulses (1-20), the peak value of the excitatory postsynaptic current (EPSC) output by the device was continuously varied within the range of 0-250 nA. The peak current was normalized and discretized into a 16-level grade code (4-bit encoding, numbered 0-15). Calibration procedure: At each wavelength, standard optical pulse sequences of 1, 2, 3, ..., 20 pulses were applied 10 times each. The peak current was collected and normalized, sorted by amplitude, and divided into 16 threshold intervals using equal intervals. At a wavelength of 350 nm, grade code 0 corresponds to a normalized peak value of 0-0.0625, and grade code 15 corresponds to 0.9375-1. At wavelengths of 450 nm and 500 nm, the threshold intervals are similar, but the absolute values ​​of the peak currents are different, reflecting the wavelength-selective response.

[0058] The multi-level state distribution varies across different wavelengths: at 350 nm, the peak current is highest (approximately 250 nA), and the state spacing is largest (approximately 15 nA / level); at 450 nm, the peak current is moderate (approximately 180 nA), and the state spacing is approximately 11 nA / level; at 500 nm, the peak current is lowest (approximately 120 nA), and the state spacing is approximately 7 nA / level. This difference allows the "wavelength dimension" to be combined with the "time dimension (pulse width, frequency, number, and interval)" to form key constraints, expanding the key space. Key space estimation: 3 wavelengths × 3 pulse widths (100 ms, 500 ms, 1000 ms) × 10 pulse numbers (1-10) × 3 pulse frequencies (1 Hz, 10 Hz, 100 Hz) × 6 wavelength application orders (3! = 6) = 5400 combinations, far exceeding the traditional single-parameter key space (approximately 100 combinations).

[0059] like Figure 14As shown, the grade code sequence {k(t)} is input into the reservoir computing unit. After the reservoir state x(t) undergoes nonlinear mapping through 300 nodes, the output state vector y(t) is determined. The classification decision unit performs binary classification on y(t): when the preset optical key conditions are met (e.g., wavelength sequence 350 nm-450 nm-500 nm, pulse interval 500 ms, pulse number 5 for each), y(t) > threshold (e.g., 0.8), and a valid decoding result (e.g., label "correct") is output; when not met, y(t) < threshold, and a rejection state (e.g., label "locked") is output.

[0060] The decoding result or grade code is mapped to PDLC drive parameters. Segmented mapping rule: lower grade codes (0-7) are mapped to the threshold range (13-18 V), and the drive voltage... V; High-level codes (levels 8-15) are mapped to the saturation range (37-42 V). V. The driving signal is a square wave pulse with a frequency of 1 kHz, a duty cycle of 50%, and an amplitude of... .

[0061] like Figure 15 As shown, the transmittance of the PDLC display device varies under different driving voltages: at 0 V, the transmittance is approximately 10% (low transmittance / high scattering state), and the pattern is not visible; at a driving voltage of 13 V (threshold voltage V), the transmittance is reduced to approximately 10%. th At a driving voltage of 18 V (the upper limit of the threshold range), the transmittance is approximately 30%; at a driving voltage of 37 V (the saturation voltage V), the transmittance is approximately 50%; at a driving voltage of 18 V (the upper limit of the threshold range), the transmittance is approximately 50%; at a driving voltage of 37 ... sat At a driving voltage of 42 V (the upper limit of the saturation range), the transmittance is approximately 70%; at a driving voltage of 42 V (the upper limit of the saturation range), the transmittance is approximately 72%.

[0062] The advantages of the segmented mapping strategy are: low-level codes are mapped to the threshold range, utilizing the steep area of ​​the transmittance-voltage curve to enhance contrast (transmittance change of approximately 40% / 5 V, or 8% / V); high-level codes are mapped to the saturation range, utilizing the flat area of ​​the transmittance-voltage curve to reduce brightness fluctuations (transmittance change of approximately 2% / 5 V, or 0.4% / V), thereby improving pattern readability and noise resistance.

[0063] like Figure 16As shown, in a typical scenario, an external target image (such as a handwritten digit "8") is converted into a light stimulation sequence that meets preset key conditions: wavelength sequence 350 nm-450 nm-500 nm, pulse interval 500 ms, 5 pulses of each wavelength, and pulse width 500 ms. The photosynaptic device outputs a time-series current response, which is converted into a voltage signal by signal conversion unit 2 and discretized into a level code sequence {k(t)} by state generation unit. The level code sequence is input into the reservoir calculation unit, which outputs a state vector y(t). The classification decision unit classifies y(t) and outputs the recognition result "8". The recognition result is mapped to the PDLC driving voltage (e.g., level code 12 corresponds to...). The PDLC display device outputs a corresponding brightness pattern to display the number "8".

[0064] The decoding and decision module performs key consistency verification. If the preset optical key conditions are met, a valid decoding result is output, and entry into the display mapping process is allowed; otherwise, a rejection status is output, and a secure output policy is triggered. The preset optical key conditions include: wavelength combination (e.g., 350 nm-450 nm-500 nm), pulse width (500 ms), number of pulses (5 each), pulse frequency (2 Hz, corresponding to a 500 ms interval), and wavelength application order (350 nm→450 nm→500 nm). Any mismatch in these parameters (e.g., changing the wavelength order to 450 nm→350 nm→500 nm, or changing the pulse interval to 200 ms) triggers a rejection status.

[0065] Secure output strategies include: First, the PDLC display device maintains 0V drive, outputting a blank pattern (with approximately 10% transmittance, the pattern is invisible); second, outputting a preset "locked" pattern (achieved through patterned ITO electrodes); third, outputting a pseudo-random brightness pattern (the drive voltage randomly varies within the range of 13-42V, making it impossible to form a stable readable pattern); and fourth, outputting a fixed watermark (such as "Unauthorized Access") to indicate unauthorized input. Through these methods, even if an attacker intercepts some optical parameters (such as wavelength) or electrical signals (such as security codes), it is difficult to obtain usable plaintext information or a stable visual output.

[0066] To ensure repeatability under different devices and environmental conditions, the state generation unit performs a calibration procedure before operation. The calibration procedure involves applying pulses with a width of 500 ms and an intensity of 1 mW / cm² at wavelengths of 350 nm, 450 nm, and 500 nm, respectively. 2 Each standard optical pulse was tested 20 times, and the peak current was collected. and normalized to (The maximum value from 20 samples). Sort by amplitude from smallest to largest, and divide by equal intervals to generate 16 threshold intervals: [0, 1 / 16], [1 / 16, 2 / 16], ..., [15 / 16, 1]. Establish a " - A "level code" mapping table. After calibration, the mapping table is stored in the memory of the state generation unit for subsequent real-time quantization encoding.

[0067] To reduce the impact of device drift on threshold stability, a reference baseline and an online correction mechanism are introduced. Reference baseline: Before each test, the baseline current is acquired in the dark. (Approximately 1 nA), used as a reference value for baseline correction. Online correction mechanism: After every 10 tests, the peak current is re-acquired and the mapping table is updated. A sliding window averaging method (window size 10) is used to reduce the impact of random noise and maintain the long-term distinguishability and consistency of the grade code.

[0068] To stably map the gradation codes to PDLC brightness output, a segmented mapping table of "gradation code-driving voltage" is established. Segmentation strategy: Low gradation codes (0-7) are mapped to the threshold range (13-18 V), utilizing the steep region of the transmittance-voltage curve to enhance contrast; high gradation codes (8-15) are mapped to the saturation range (37-42 V), utilizing the flat region of the transmittance-voltage curve to reduce brightness fluctuations. Driving parameters: Square wave pulse, frequency 1 kHz (corresponding to a period of 1 ms), duty cycle 50% (high level time 0.5 ms, low level time 0.5 ms), amplitude... It is determined by the grade code k.

[0069] To improve decoding robustness, signal conversion unit 2 performs filtering on the synaptic response (second-order low-pass filter, cutoff frequency 1kHz, attenuating high-frequency noise) and baseline correction (subtracting the dark-state baseline voltage). ) and normalization (divided by peak voltage) Feature extraction preferably uses relative feature quantities: the pairwise pulse facilitated (PPF) ratio A2 / A1 (reflecting short-time memory characteristics and unaffected by absolute current value), and the peak increment (A1-). (Eliminating baseline drift) and normalizing the decay time constant (Using the reference decay time to eliminate device variability). Relative characteristic quantities reduce the impact of device variability and slow drift on the decoding results, improving system stability.

[0070] In one optional embodiment, the photostimulation response unit 1 includes a first synaptic channel and a second synaptic channel in parallel. The first synaptic channel adopts an HfAlO / NbOx heterostructure, responding to wavelengths of 350 nm and 450 nm, and is used to encode color information; the second synaptic channel adopts a single NbOx layer (without HfAlO), responding to a wavelength of 500 nm, and is used to encode luminance information. The outputs of the two channels are fused (weighted summation or concatenation) and then input into the decoding and decision module to expand the encoding dimension and key space. The weights of the first and second channels can be adjusted according to the application scenario: in anti-counterfeiting authentication scenarios, color information has a higher weight (e.g., 0.7); in luminance modulation scenarios, luminance information has a higher weight (e.g., 0.7).

[0071] To verify the necessity and superiority of the optimal scheme for the PDLC material system, three groups of PDLC samples with different formulations (A, B, and C) were set up to compare their microstructure and electro-optic properties (threshold voltage, saturation voltage, contrast, and response time).

[0072] Group A: Baseline formulation.

[0073] Group A PDLC samples used a baseline formulation of liquid crystal phase E7 (70 wt%) and crosslinking agent UV6300 (28 wt%), photoinitiator Irgacure 651 (2 wt%), without adding acrylic monomers or nanoparticles. The preparation process was the same as step A5 in Example 1, with a box thickness of 10 μm and a UV curing time of 8 minutes.

[0074] like Figure 5 As shown, the scanning electron microscope (SEM) morphology image (magnification 5000×) of group A samples reveals that the liquid crystal droplets are approximately 5-10 μm in size and unevenly distributed, with droplet aggregation in some areas. Transmittance-voltage curves (test conditions: room temperature 25 °C, square wave pulse 1 kHz, duty cycle 50%): transmittance is approximately 10% at 0 V (low transmittance / high scattering state), threshold voltage V... th Approximately 9 V (voltage at which transmittance reaches 30%), saturation voltage V sat At approximately 27 V (voltage at which 70% transmittance is achieved), the saturation transmittance is approximately 72%, and the contrast ratio (saturation transmittance / 0 V transmittance) is approximately 7.2. Response time (test conditions: drive voltage step from 0 V to 27 V): rise time (10%-90% transmittance) is approximately 65 ms, and fall time (90%-10% transmittance) is approximately 70 ms.

[0075] Group A formulation serves as the baseline, containing only a crosslinking agent. The liquid crystal droplets are large in size and unevenly distributed, resulting in a lower threshold voltage, lower contrast, slower response time, and generally mediocre performance.

[0076] Group B: Add acrylic monomers.

[0077] Group B PDLC samples were prepared by adding THFMA acrylic monomers to the preparation of Group A samples. The formulation was as follows: liquid crystal phase E7 (70 wt%), crosslinking agent UV6300 (18 wt%), THFMA (10 wt%), and photoinitiator Irgacure 651 (2 wt%). The preparation process was the same as step A5 in Example 1.

[0078] like Figure 6 As shown, the scanning electron microscope (SEM) morphology images of group B samples reveal that the liquid crystal droplet size has decreased to 3-6 μm and is more uniformly distributed, with reduced droplet aggregation. Transmittance-voltage curve: Threshold voltage V th Approximately 11 V, saturation voltage V sat Approximately 32 V, contrast ratio approximately 10. Response time: rise time approximately 50 ms, fall time approximately 55 ms.

[0079] The addition of acrylic monomers increases crosslinking density, resulting in a denser polymer network, smaller liquid crystal droplet size, and more uniform distribution. The smaller droplet size leads to enhanced scattering and an increased threshold voltage (from 9 V to 11 V); the uniform droplet distribution increases the steepness of the transmittance-voltage curve, improving contrast (from 7.2 to 10); and the denser polymer network accelerates the reorientation of liquid crystal molecules, shortening the response time (rise time decreases from 65 ms to 50 ms).

[0080] Group C: Acrylic monomers and nanoparticles added.

[0081] The PDLC samples in Group C were prepared by adding AZO nanoparticles to the samples in Group B. The formulation was as follows: liquid crystal phase E7 (70 wt%), crosslinking agent UV6300 (15 wt%), THFMA (10 wt%), AZO nanoparticles (3 wt%), and photoinitiator Irgacure 651 (2 wt%). The preparation process was the same as step A5 in Example 1.

[0082] like Figure 7 As shown, the scanning electron microscope (SEM) morphology images of group C samples reveal that the liquid crystal droplet size has further decreased to 2-4 μm and is most uniformly distributed, with a more regular droplet shape (approaching spherical). Transmittance-voltage curve: Threshold voltage V th Approximately 13 V, saturation voltage V sat Approximately 37 V, contrast ratio approximately 12. Response time: rise time approximately 40 ms, fall time approximately 45 ms.

[0083] The addition of nanoparticles, with AZO nanoparticles (dielectric constant approximately 10) dispersed within the polymer network, increases the dielectric constant of the composite material, leading to a further increase in the threshold voltage (from 11 V to 13 V). As nucleating agents, the nanoparticles promote the formation of liquid crystal droplets, resulting in smaller, more uniformly distributed, and more regularly shaped droplets, thus improving scattering efficiency and further enhancing contrast (from 10 to 12). The interfacial interaction between the nanoparticles and the polymer network enhances network rigidity, further accelerating the reorientation rate of liquid crystal molecules and significantly shortening the response time (rise time reduced from 50 ms to 40 ms).

[0084] As shown in Table 1, the formulation in group C (containing crosslinking agent + acrylic monomer + nanoparticles) exhibits the highest threshold voltage (13V), highest saturation voltage (37V), highest contrast ratio (12), and fastest response time (rise time 40 ms, fall time 45ms), making it the preferred embodiment of this invention. By controlling the crosslinking density, liquid crystal droplet size, and dielectric properties through component engineering (adding acrylic monomer and nanoparticles), the threshold voltage of the PDLC was improved (by 44%, from 9V to 13V), contrast ratio (by 67%, from 7.2 to 12), response speed (by 38%, with rise time decreasing from 65 ms to 40 ms), and stability.

[0085]

[0086] It should be further noted that the structural composition, material system, process parameters, signal processing flow, and encoding / mapping rules described in the above embodiments are all illustrative examples of preferred embodiments of the present invention, intended to facilitate understanding and implementation of the present invention by those skilled in the art, and should not be construed as limiting the scope of protection of the present invention. Without departing from the concept and technical essence of the present invention, those skilled in the art can, according to actual application needs, adjust the thickness range of each layer of material (e.g., HfAlO thickness 5-20 nm, NbOx thickness 10-30 nm), deposition method (e.g., ALD, PVD, solution method), and pulse optical parameters (e.g., wavelength 200-600 nm, pulse width 10 ms-10 s, light intensity 0.1-10 mW / cm²). 2 All technical solutions formed by equivalent substitutions, combinations, adjustments, or improvements to the following methods shall fall within the protection scope of this invention: quantization threshold setting methods (such as equal interval division, quantile division), decoding algorithm module forms (such as reservoir calculation, recurrent neural network, convolutional neural network), PDLC formulations (such as liquid crystal phase type, crosslinking agent type, nanoparticle type and concentration) and driving parameter mapping relationships (such as linear mapping, piecewise mapping, lookup table mapping).

[0087] In summary, the neuromorphic encryption decoding and brightness modulation display device of the present invention also includes the following effects: Integrated closed loop: Multispectral input, state formation, quantization encoding, decoding decision and visual output are realized within this device, reducing external links and leakage risks, and reducing power consumption and latency; Multi-dimensional key: Combining wavelength combinations with pulse / timing parameters to form optical key constraints expands the key space and improves resistance to theft and false triggering. Unified Interface: The grade code is used as a unified interface for encoding and display driving, which reduces the dependence on high-precision ADC and complex software decoding, and improves feasibility and consistency. Display robustness: Segmented driving mapping is established using the PDLC threshold range and saturation range to improve contrast and enhance brightness output stability and readability; Process compatibility: It adopts mature thin film deposition, heat treatment, UV curing and encapsulation processes, and the photoelectric synapses can be arrayed and the PDLC electrodes can be patterned / partitioned, which facilitates system expansion.

[0088] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A neuromorphic encryption decoding and brightness modulation display device, characterized in that, It includes a light stimulus response unit, a signal conversion unit, a state generation unit, and a visual feedback unit; The photostimulation response unit includes photosynaptic devices for generating a history-dependent electrical response under illumination by light pulses of different wavelengths; The signal conversion unit is electrically connected to the photostimulation response unit and is used to read out and convert the electrical response to obtain a voltage signal or time series signal that can be used for state characterization. The state generation unit is connected to the signal conversion unit and is used to discretize voltage signals or time series signals into multi-level state codes. The visual feedback unit is signal-connected to the state generation unit. The visual feedback unit includes a PDLC display device and a driving circuit. It is used to generate corresponding driving parameters according to multi-level status codes and control the output transmittance change, grayscale pattern and status prompt of the PDLC display device to realize the neuromorphic response closed loop from light input to visual feedback output.

2. The neuromorphic encryption decoding and brightness modulation display device as described in claim 1, characterized in that, The photoelectric synaptic device adopts an HfAlO / NbOx heterostructure. The HfAlO trap control layer and the NbOx photosensitive layer form a heterostructure interface. The area near the heterostructure interface contains oxygen vacancy-related defect trap states or interface trap states to achieve the capture and release of photogenerated carriers, resulting in short-term and long-term plasticity.

3. The neuromorphic encryption decoding and brightness modulation display device as described in claim 1 or 2, characterized in that, The photoelectric synaptic device generates a wavelength-selective electrical response to light stimulation in the 250-500nm band, and forms more than 16 discrete conductive states under excitation at least two wavelengths in 350nm, 450nm and 500nm. The state generation unit generates multi-level state codes based on at least one of the following: peak value, steady-state value, integral value, decay time constant, and pulse sequence statistical characteristics.

4. The neuromorphic encryption decoding and brightness modulation display device as described in claim 1, characterized in that, The preset optical key conditions include at least one of the following: wavelength combination, pulse width, number of pulses, pulse frequency, pulse interval, and wavelength application order; The photostimulation response unit includes a first synaptic channel and a second synaptic channel in parallel. The first synaptic channel is used to encode color information, and the second synaptic channel is used to encode structural information, brightness information, or watermark information, forming a hierarchical encoding or multi-factor key encoding. Output a valid decoding result when the preset optical key conditions are met, and output a rejection status when the preset optical key conditions are not met.

5. The neuromorphic encryption decoding and brightness modulation display device as described in claim 1, characterized in that, The signal conversion unit includes a transimpedance amplifier circuit and a charge amplifier circuit, and also includes a baseline correction unit, a normalization unit, and a noise suppression unit; The state generation unit includes a discretization mapping unit, which is used to discretize the electrical response into a multi-level rating code and map it into PDLC driving parameters to form a token interface that can directly drive the display. The visual feedback unit also includes a decoding module, which is signal-connected to the state generation unit. The decoding module includes a reservoir calculation unit and a classification decision unit for outputting the decoding results.

6. The neuromorphic encryption decoding and brightness modulation display device as described in claim 1, characterized in that, The PDLC display device includes a PDLC thin film layer sandwiched between two transparent ITO electrodes. The driving signal output by the driving circuit is an AC driving signal or a pulse driving signal, which enables the PDLC display device to have a threshold range and a saturation range and to achieve continuous brightness modulation or graded grayscale output. The ITO electrode is a patterned electrode or a partitioned electrode; the PDLC thin film layer is a liquid crystal microdroplet-polymer network structure, and the polymerizable monomer system includes at least one of a crosslinking agent, an acrylic monomer, and doped nanoparticles.

7. A neuromorphic encryption decoding and brightness modulation display method, characterized in that, Using the neuromorphic encryption decoding and brightness modulation display device according to any one of claims 1-6, Includes the following processes: Set preset optical key conditions and map the information to be processed into a multi-wavelength optical pulse sequence; Apply the optical pulse sequence and obtain the electrical response; The electrical response is read out, conditioned, and quantized into a grade code or feature vector to generate an encoding result; The encoding result is decoded when the preset optical key condition is met to obtain the decoding result; The encoding or decoding result is mapped to a PDLC drive signal and loaded onto the PDLC display device to output a brightness pattern; When the preset optical key conditions are not met, a rejection state is output and the PDLC display device is controlled to output a rejection pattern or a blank pattern.

8. The neuromorphic encryption decoding and brightness modulation display method as described in claim 7, characterized in that, Decoding involves calculating the state vector through a reservoir and classifying or reconstructing the state vector using a classification decision unit; the preset optical key conditions include at least a joint constraint on the wavelength combination and the order of wavelengths. The PDLC drive signal is obtained by mapping a level code or state vector, and the mapping parameters include at least the drive voltage amplitude level; brightness enhancement, brightness compression, or graded grayscale output are achieved by configuring the drive operating point near the threshold range or near the saturation range.

9. A method for manufacturing the neuromorphic encryption decoding and brightness modulation display device according to any one of claims 1-6, characterized in that, Includes the following processes: A substrate is provided and cleaned to form a first electrode and a second electrode; Formation of HfAlO trap control layer; A NbOx photosensitive layer is formed to construct an HfAlO / NbOx heterostructure, and interface trap states or defect trap states are introduced near the heterostructure by annealing or oxygen partial pressure regulation during deposition. Encapsulate multispectral opto-synaptic units and connect them to various functional modules; A PDLC thin film layer was fabricated and sandwiched between two transparent ITO electrodes to form a PDLC brightness modulation display module; Establish the mapping relationship between the output of the multispectral photoelectric synapse unit and the PDLC driving signal and complete the integration.

10. The preparation method according to claim 9, characterized in that, The process of fabricating a PDLC thin film layer and sandwiching it between two transparent ITO electrodes to form a PDLC brightness modulation display module includes: A liquid crystal phase is mixed with a polymerizable monomer system and a photoinitiator is added. The mixture is then coated or infused, gap-controlled and UV-cured to form a PDLC thin film layer with a liquid crystal microdroplet-polymer network structure. The polymerizable monomer system includes at least one of a crosslinking agent, an acrylic monomer, and doped nanoparticles. Establish and integrate the mapping relationship between the multispectral photoelectric synapse unit output and the PDLC drive signal, including: A one-to-one or segmented correspondence is established between the multi-level discrete states of the multispectral photoelectric synapse unit and the PDLC driving parameters. Mapping parameters or mapping tables are established for different wavelength channels to output a target brightness pattern when the preset optical key conditions are met, and a rejection pattern or blank pattern is output when the conditions are not met.

Images