Description:FIELD OF THE INVENTION
The present invention relates to the field of medical diagnostic systems, particularly the use of artificial intelligence (AI) in dermatology. More specifically, it concerns a hybrid diagnostic framework for detecting and classifying Erythemato-Squamous Diseases (ESD) using a Harmonic Archery Algorithm (HAA) for feature selection and a Deep Spiking Attention Network (DSNAN) for classification and interpretability. The invention combines biologically inspired learning models with metaheuristic optimization techniques to provide accurate, interpretable, and resource-efficient diagnostic support suitable for both advanced clinical centers and low-resource healthcare environments.
BACKGROUND OF THE INVENTION
Erythemato-Squamous Diseases (ESD) are a class of dermatological disorders that include conditions such as psoriasis and seborrheic dermatitis, which share overlapping symptoms and clinical presentations. Diagnosis is often difficult, relying heavily on histopathology, biopsies, and subjective physician assessment. Misdiagnosis is common due to the similarity of symptoms across subtypes, resulting in delays in treatment and potential complications.
Existing diagnostic systems using conventional machine learning models such as Support Vector Machines (SVM), decision trees, and k-NN rely on hand-crafted features, which often result in poor accuracy. Deep learning models, including convolutional neural networks (CNNs), have shown promise but demand large datasets and computationally expensive resources, making them unsuitable for real-time or portable diagnostic applications. Furthermore, most current AI systems function as black boxes, offering little or no interpretability, which reduces clinician trust and acceptance.
Feature selection techniques such as genetic algorithms (GA) or particle swarm optimization (PSO) have been attempted but suffer from premature convergence, instability, and poor handling of noisy or imbalanced datasets. Spiking Neural Networks (SNNs), while biologically inspired, have rarely been combined with advanced feature optimization or attention mechanisms in medical diagnosis. No known system integrates a biologically inspired metaheuristic feature selection algorithm like Harmonic Archery with a spiking attention-based neural network for real-time, explainable diagnosis of ESD.
US20200054714A1: Provided herein are variants and fusions of fibroblast growth factor 19 (FGF19), variants and fusions of fibroblast growth factor 21 (FGF21), fusions of FGF19 and/or FGF21, and variants or fusions of FGF19 and/or FGF21 proteins and peptide sequences (and peptidomimetics), having one or more activities, such as glucose lowering activity, and methods for and uses in treatment of hyperglycemia and other disorders.
US2016340426A1: The use of a modulator of the Nav 1.9 sodium channel is described for treating an inflammatory skin disease. Also described, is a pharmaceutical composition including a modulator of the Nav 1.9 sodium channel, and in vitro diagnostic methods based on the detection or quantification of Nav 1.9.
The invention aims to overcome the limitations of existing diagnostic systems for ESD. A key objective is to provide a hybrid diagnostic framework that combines metaheuristic feature selection with biologically inspired spiking neural attention networks to enhance diagnostic accuracy and interpretability. Another objective is to enable automatic selection of the most relevant clinical and histopathological features, thereby reducing redundancy and improving efficiency.
The invention also aims to provide an explainable AI framework that mimics human cognitive processes, making diagnostic results interpretable for clinicians through attention-based explanations. Further, it seeks to achieve robust performance even on small, noisy, or class-imbalanced datasets, common in dermatological practice. Another objective is to make the system resource-efficient, so it can be deployed on portable diagnostic equipment and in rural healthcare centers. Ultimately, the invention seeks to deliver a cost-effective, scalable, and accurate diagnostic framework that enhances dermatology healthcare delivery.
SUMMARY OF THE INVENTION
This summary is provided to introduce a selection of concepts, in a simplified format, that are further described in the detailed description of the invention.
This summary is neither intended to identify key or essential inventive concepts of the invention and nor is it intended for determining the scope of the invention.
The invention discloses a hybrid diagnostic system for Erythemato-Squamous Disease detection that integrates a Harmonic Archery Algorithm (HAA) for optimal feature selection and a Deep Spiking Neuron Attention Network (DSNAN) for classification and interpretability. The HAA module dynamically identifies the most informative clinical and histopathological features, reducing redundancy and enhancing classification performance. These selected features are encoded into spike trains and processed by DSNAN, which mimics biologically inspired neuronal communication.
An attention mechanism within the DSNAN guides the classification process by emphasizing the most diagnostically relevant features, improving interpretability and trustworthiness. The hybrid framework enhances diagnostic precision, reduces computational complexity, and provides sparse, energy-efficient learning, making it suitable for deployment in both advanced and low-resource healthcare environments.
By addressing issues of misdiagnosis, poor interpretability, and computational inefficiency in current systems, the invention provides a novel, explainable, and practical AI-based diagnostic solution for ESD.
To further clarify advantages and features of the present invention, a more particular description of the invention will be rendered by reference to specific embodiments thereof, which is illustrated in the appended drawings. It is appreciated that these drawings depict only typical embodiments of the invention and are therefore not to be considered limiting of its scope. The invention will be described and explained with additional specificity and detail with the accompanying drawings.
The proposed invention illustrates a hybrid diagnosis model with a Harmonic Archery Algorithm (HAA) for feature selection and a Deep Spiking Neuron Attention Network (DSNAN) for right kind Erythemato-Squamous Disease (ESD) selection. ESD is a dermatology medical skin disease with multiple classes of dermatology with the same symptoms. The identification contributes significantly to classification accuracy, decreases diagnostic time, and incorporates low feature redundancy with the integration of nature-inspired optimization as well as biologically inspired spiking networks. The integration includes dynamic weighting adaptation of the features and improved pattern recognition utilizing clinical data. The system has been devised to be applied in the context of a clinical decision-support system to counsel dermatologists and physicians with real-time, AI-based evidence-based suggestions to diagnose diseases in the right manner and estimate their severity
BRIEF DESCRIPTION OF THE DRAWINGS
The illustrated embodiments of the subject matter will be understood by reference to the drawings, wherein like parts are designated by like numerals throughout. The following description is intended only by way of example, and simply illustrates certain selected embodiments of devices, systems, and methods that are consistent with the subject matter as claimed herein, wherein:
FIGURE 1: SYSTEM ARCHITECTURE
The figures depict embodiments of the present subject matter for the purposes of illustration only. A person skilled in the art will easily recognize from the following description that alternative embodiments of the structures and methods illustrated herein may be employed without departing from the principles of the disclosure described herein.
DETAILED DESCRIPTION OF THE INVENTION
The detailed description of various exemplary embodiments of the disclosure is described herein with reference to the accompanying drawings. It should be noted that the embodiments are described herein in such details as to clearly communicate the disclosure. However, the amount of details provided herein is not intended to limit the anticipated variations of embodiments; on the contrary, the intention is to cover all modifications, equivalents, and alternatives falling within the scope of the present disclosure as defined by the appended claims.
It is also to be understood that various arrangements may be devised that, although not explicitly described or shown herein, embody the principles of the present disclosure. Moreover, all statements herein reciting principles, aspects, and embodiments of the present disclosure, as well as specific examples, are intended to encompass equivalents thereof.
The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments. As used herein, the singular forms “a",” “an” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms “comprises,” “comprising,” “includes” and/or “including,” when used herein, specify the presence of stated features, integers, steps, operations, elements and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and/or groups thereof.
It should also be noted that in some alternative implementations, the functions/acts noted may occur out of the order noted in the figures. For example, two figures shown in succession may, in fact, be executed concurrently or may sometimes be executed in the reverse order, depending upon the functionality/acts involved.
In addition, the descriptions of "first", "second", “third”, and the like in the present invention are used for the purpose of description only, and are not to be construed as indicating or implying their relative importance or implicitly indicating the number of technical features indicated. Thus, features defining "first" and "second" may include at least one of the features, either explicitly or implicitly.
Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which example embodiments belong. It will be further understood that terms, e.g., those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the relevant art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein.
The proposed invention illustrates a hybrid diagnosis model with a Harmonic Archery Algorithm (HAA) for feature selection and a Deep Spiking Neuron Attention Network (DSNAN) for right kind Erythemato-Squamous Disease (ESD) selection. ESD is a dermatology medical skin disease with multiple classes of dermatology with the same symptoms. The identification contributes significantly to classification accuracy, decreases diagnostic time, and incorporates low feature redundancy with the integration of nature-inspired optimization as well as biologically inspired spiking networks. The integration includes dynamic weighting adaptation of the features and improved pattern recognition utilizing clinical data. The system has been devised to be applied in the context of a clinical decision-support system to counsel dermatologists and physicians with real-time, AI-based evidence-based suggestions to diagnose diseases in the right manner and estimate their severity.
The proposed invention provides an intelligent diagnostic framework designed for the accurate detection of Erythemato-Squamous Diseases. It integrates a novel Harmonic Archery Algorithm (HAA) for feature selection with a Deep Spiking Neuron Attention Network (DSNAN) for classification, combining optimization efficiency with biologically inspired computation.
The diagnostic process begins with data input, which may include clinical and histopathological features collected from diagnostic laboratories or electronic health records. These raw inputs are often heterogeneous, noisy, and incomplete, requiring preprocessing before analysis. The preprocessing stage normalizes the data, handles missing values, and ensures consistent scaling across features to prepare the dataset for effective feature selection.
The Harmonic Archery Algorithm (HAA) serves as the core feature selector in the invention. Inspired by harmonic motion and archery dynamics, the algorithm identifies the most informative and non-redundant features from high-dimensional datasets. This reduces computational burden while enhancing diagnostic performance, especially in small or imbalanced datasets typical of dermatological studies.
Once features are selected, they are encoded into temporal spike trains, which serve as inputs for the Deep Spiking Neuron Attention Network (DSNAN). Spiking neurons simulate biologically plausible computation by transmitting information as time-dependent spike events rather than continuous activations, thus providing energy-efficient and sparse representations.
The DSNAN incorporates an attention mechanism that dynamically assigns weights to the most diagnostically significant features during the decision-making process. This attention-driven learning improves classification accuracy while also producing interpretability tools, such as attention maps, that help clinicians understand the rationale behind predictions.
The classification layer of the DSNAN predicts the presence or absence of ESD, as well as the specific subtype, along with confidence scores. This provides a reliable diagnostic output that can be validated against clinical expertise. Importantly, the interpretability mechanisms ensure that physicians can visualize why certain features contributed to the diagnosis.
The invention further includes a user-friendly interface where clinicians can input patient data and receive diagnostic predictions in real time. Results are displayed with confidence scores and explanations, enabling evidence-based decision-making. A logging and storage system securely stores both inputs and outputs for future reference and continuous model improvement.
The framework supports continuous learning by updating its parameters incrementally as new data is introduced. This ensures that the diagnostic engine remains adaptable and improves performance over time, reflecting real-world variations in disease patterns.
In terms of deployment, the invention is versatile. It can be hosted on cloud infrastructures for large-scale use in hospitals, deployed locally on institutional servers, or embedded in portable edge devices for rural healthcare delivery. Low power consumption of spiking computation further supports deployment in resource-constrained environments.
Security and privacy of patient data are also ensured. The system employs AES-256 encryption for sensitive inputs and outputs, with role-based access control for clinicians and researchers. Secure communication protocols like HTTPS and MQTT are used for data transmission in both cloud and IoT-based deployments.
Alternative embodiments of the invention include replacing HAA with other metaheuristic algorithms, such as Whale Optimization or Firefly Algorithm, to explore comparative performance. Similarly, while DSNAN is the preferred model, hybrid neural networks like BiLSTM or CNN-RNN with attention layers can be used if hardware constraints exist.
The system is also capable of multimodal data integration. In addition to clinical and histopathological features, the framework can incorporate dermoscopic or microscopic images for richer diagnostic capability. This multimodal approach further enhances accuracy and robustness.
An optional severity scoring module can also be integrated into the framework. This regression-based layer provides quantitative severity levels, enabling monitoring of disease progression and treatment response over time.
The invention’s adaptability makes it suitable for various clinical settings, including tele-dermatology platforms where remote diagnosis is critical. Its energy efficiency, scalability, and interpretability make it ideal for bridging the gap between AI models and clinical decision-making.
Overall, the invention introduces a transformative diagnostic solution for ESD, combining metaheuristic optimization, biologically inspired spiking networks, and attention-driven interpretability to deliver accurate, efficient, and clinically deployable diagnostic support.
KEY COMPONENTS
1. Data Preprocessing Unit
Normalization and histopathological and clinical data cleaning.
Handling missing values, outliers, and feature scaling.
2. Feature Selector based on Harmonic Archery Algorithm (HAA)
A bio-inspired optimization module for most significant feature selection from the dataset.
Very high feature importance and low redundancy.
3. Deep Spiking Neuron Attention Network (DSNAN)
Consisting of spiking neurons processing selected features with time-dependent spike signals.
Decision-making high-impact feature attention mechanism.
4.Training & Evaluation Engine
o Tests hybrid model on labeled data with proper cross-validation.
o Checks model's accuracy, sensitivity, specificity, and F1-score.
5.User Interface Dashboard
Allows clinicians to enter patient data and receive diagnostic predictions.
Allows disease classification and probability scores.
6. Cloud Storage & Logging System
Secures patient input and diagnostic output storage.
Allows retraining model and versioning.
All of them ensure secure, interpretable, and real-time ESD diagnosis with scalability to deploy in other healthcare centers.
TECHNOLOGY USED
• Sensors (optional for real-time skin image capture):
Dermatoscopic high-resolution sensors or smartphone camera integrations for feature images.
• Communication Protocols
HTTP/HTTPS for secure data transmission.
MQTT for low-transmission in IoT-based deployments.
•Power Supply:
Desktop/Cloud-deployed: Mains power supply.
Embedded deployments: Rechargeable lithium-ion battery or solar panels for rural deployments.
•_Software Algorithms:
Harmonic Archery Algorithm (HAA): New, nature-inspired algorithm using harmonic motion and archery to select the best subset of features.
Spiking Neural Network: Using leaky integrate-and-fire (LIF) neuron models for temporal processing.
Attention Mechanism: Dynamically scales classification-time feature weights to mimic attentive human cognition.
•Frameworks & Tools:
Python, PyTorch, BindsNET for SNN simulation.
Scikit-learn for baseline and pre-processing.
TensorBoard or Matplotlib for monitoring performance.
•Security & Privacy:
AES-256 encryption of patient data.
Role-based access control for clinicians and physicians.
•Deployment Environments:
Cloud (AWS/GCP) for large-scale deployment.
Local servers for hospital IT infrastructure.
Portable AI chip-based edge devices for bedside diagnostics.
Step-by-Step Working Functionality
The suggested system is a linear pipeline from input data to diagnostic output. Step by step, the process works as follows:
1. Data Input and Gathering
Clinical and histopathological features of suspected Erythemato-Squamous Disease (ESD) patients are collected from diagnostic laboratories or electronic health records.
2. Data Preprocessing
Raw data is normalized, missing values filled up, and noise removal. Features are unit-normalized to same order to match learning model.
3. Harmonic Archery Algorithm (HAA) Feature Selection
HAA module simulates harmonic motion and archery physics in trying to discover the best set of features from feature set. This is for dimension reduction and signal-to-noise ratio enhancement.
4. Spiking Neural Network Encoding of Target Features
Target features are encoded as temporal spike trains for DSNN processing.
5. Duplex Processing via Deep Spiking Neuron Attention Network
DSNN model with attention mechanism is used for spike train processing. The attention layer identifies the most important input signals, enhancing interpretability as well as classification accuracy.
6. Classification and Prediction
Prediction of ultimately presence or absence of ESD and subtype, along with confidence score, is performed in the output layer.
7. Display of Results
Results are given in straightforward user-friendly style. Input characteristics, classification output, and distribution of attention weight are given in summary for better clinical interpretation.
8. Continuous Learning and Logging of Data
Every case is logged to render the model enriched with incremental step-by-step learning with the progression of time. The model is enhanced with the passage of time through incremental learning and novel patterns.
Project-oriented approach yields cost-effective, explainable, and precise diagnosis of ESD.
Accurate Functionality
Accurate functioning of the aforesaid system here is particularly detecting and diagnosing subtypes of Erythemato-Squamous Disease based on a hybrid, intelligent AI system. This system achieves:
•maximum extraction of clinical/histopathological features with Harmonic Archery Algorithm.
•encodes on selected information spikes for biologically inspired processing.
•employs a Stress-Aware Deep Spiking Neuron Network to emphasize crucial diagnostic features.
•generate a subtype marker and confidence estimate diagnostic prediction.
•Provide interpretability tools (e.g., attention maps) to doctors.
•Supports real-time decision-making in online and offline medical settings.
ADVANTAGES OF THE INVENTION
• Environment: Spiking neural networks application reduces the computation resources and power required and thereby reduces AI solutions as environment harmful.
• Society: Facilitates early skin disease diagnosis at low cost, especially for remote regions. Reduces the level of invasive diagnosis and improves the quality of treatment.
• Country: Enhances country health infrastructure via scaling AI-based diagnosis. Complements rural health missions and electronic health initiatives towards better dermatology access.
NOVELTY:
Novelty literally refers to synesthetic union of Harmonic Archery Algorithm and Deep Spiking Attention Mechanisms. Unlike hand-designed feature fixed choice of features in traditional machine learning and dense neural networks, our method emulates sparse spike-based human brain learning and harmonically tunes features in accordance with behavior-inspired metaheuristics. Unlike existing approaches that are strongly dependent on hand-designed features or CNNs, our method dynamically selects the useful clinical and histopathological features selectively to aid the robustness of small and class-imbalanced classification. The attention layer also guides pathological marker learning by runtime adaptation of the weighting on a spike-event-based form that supports interpretability and performance enhancement. The process, which borrowed from the biological one, also makes the framework better suited to low-resource clinical environments.
, C , Claims:1. An intelligent diagnostic framework for Erythemato-Squamous Disease detection, comprising a Harmonic Archery Algorithm-based feature selector and a Deep Spiking Neuron Attention Network, wherein the feature selector identifies significant clinical and histopathological features and the spiking attention network classifies disease subtypes with interpretability.
2. The framework as claimed in claim 1, wherein the Harmonic Archery Algorithm reduces redundancy and improves classification accuracy by selecting an optimal subset of features from high-dimensional dermatological datasets.
3. The framework as claimed in claim 1, wherein the Deep Spiking Neuron Attention Network processes encoded temporal spike trains to emulate biologically inspired sparse computation.
4. The framework as claimed in claim 1, wherein the attention mechanism assigns dynamic weights to significant diagnostic features, enhancing interpretability and clinical trust.
5. The framework as claimed in claim 1, wherein the system provides real-time predictions with subtype classification and associated confidence scores.
6. The framework as claimed in claim 1, wherein the diagnostic results are presented through an interactive interface with visual explanations such as attention maps.
7. The framework as claimed in claim 1, wherein the system incorporates AES-256 encryption and role-based access control for secure handling of patient data.
8. The framework as claimed in claim 1, wherein the framework supports deployment in cloud, local server, and portable edge-device environments.
9. The framework as claimed in claim 1, wherein the system is capable of incremental learning to adapt to new patient data over time.
10. The framework as claimed in claim 1, wherein multimodal inputs, including dermoscopic or microscopic images, are integrated with clinical features for enhanced diagnostic accuracy.