An integrated system for non-invasive subcutaneous vein detection and vascular abnormality diagnosis
Patent Information
- Application Number
- IN202521058719
- Authority / Receiving Office
- IN · IN
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2025-06-19
- Publication Date
- 2026-09-30
AI Technical Summary
Existing vein detection and vascular abnormality diagnosis technologies lack real-time dynamic diagnostic capabilities, integration of near-infrared thermographic imaging, and deep learning-based interpretation, limiting their effectiveness in procedural guidance and early disease detection.
An integrated system combining near-infrared thermographic imaging with deep learning algorithms, utilizing a NIR thermographic sensor array, high-resolution imaging camera, and deep learning processing to capture, process, and display real-time vascular patterns and anomalies, enhancing vein visibility and diagnostic accuracy.
Enables real-time, precise vein detection and early vascular abnormality diagnosis, improving procedural efficiency and patient comfort by providing immediate clinical feedback and adaptive learning capabilities.
Abstract
Description
TECHNICAL FIELD OF INVENTIONThe present invention is related to the field of electronics engineering. Morespecifically, it relates to an integrated system for non-invasive subcutaneous veindetection and vascular abnormality diagnosis, configured to operate in real-time andenhance clinical interventions.BACKGROUND OF THE INVENTIONThe background information herein below relates to the present disclosure butis not necessarily prior art.US20150112260A1 disclosed herein systems and methods are provided fornon-invasive detection of blood vessels. The systems and methods cool uniformly atissue volume below a skin region for a specified cooling period and then imagevessel thermal footprints of vessels below the skin as they heat up the skin region. Thesystems comprise a thermal imaging device configured to image the skin region afterthe cooling period, an image processor arranged to identify, in images captured by thethermal imaging device, which arise on the skin region after discontinuation of thecooling, and displaying means configured to present the identified vessel thermalfootprints. The system and methods may analyze the spatio-temporal patterns of thenatural heating of the skin surface to derive data on the location of the vessels underthe skin. Three-dimensional (3D) imaging optics and techniques may further enhancethe vessel imaging.US10751486B2 this patent describes a system utilizing infrared lightabsorption to locate veins and assist in patient identification. It employs infraredimaging to enhance vein visibility, aiding healthcare professionals during procedureslike injections. But these systems, however, lack dynamic diagnostic capabilities anddo not incorporate thermal or deep learning-based interpretation.US8463364B2 this patent describes a portable vein viewer apparatus may bebattery powered and hand-held to reveal patient vasculature information to aid invenipuncture processes. the apparatus comprises a first laser diode emitting infraredlight, and a second laser diode emitting only visible wavelengths, wherein vasculatureabsorbs a portion of the infrared light causing reflection of a contrasted infraredimage. But they contribute to non-invasive vascular health monitoring, they areprimarily static and diagnostic in focus, without real-time procedural guidance.US8244333B2 this patent describes a miniature vein enhancer that includes aminiature projection head. the miniature projection head may be operated in one ofthree modes, AFM, DBM, and RTM. the miniature projection head of the presentinvention projects an image of the veins of a patient, which aids the practitioner inpinpointing a vein for an intravenous drip, blood test, and the like. the miniatureprojection head may have a cavity for a power source or it may have a power sourcelocated in a body portion of the miniature vein enhancer. but in this invention lackanalytical depth, machine learning integration, or the ability to detect vascularanomalies or systemic conditions.KR102176196B1 this patent describes an apparatus and method for real-timesuperficial vein detection, which detects the shape of superficial veins by analyzingthe reaction specificity of hemoglobin in the blood in a specific wavelength range in anear-infrared (NIR) irradiation environment, and projects the shape in real time on theoriginal skin surface where blood vessels are located by applying a high-speed imageprocessing algorithm. But it does not incorporate diagnostic algorithms capable ofidentifying vascular anomalies or early signs of disease.Hence, none of the above technologies fully address the integration of near-infrared thermographic imaging, real-time deep learning-based segmentation, anddiagnostic feedback within a portable clinical system. The present inventionovercomes these limitations by combining thermographic sensing, real-time imageprocessing, and artificial intelligence to guide intravenous access and enable early-stage diagnosis of vascular abnormalities, offering both procedural aid and predictivediagnostic functionality.OBJECTIVE OF THE INVENTIONThe primary objective of the present invention is to provide an integratedsystem for non-invasive subcutaneous vein detection and vascular abnormalitydiagnosis.Another objective of the invention is to develop a non-invasive, integratedimaging system that combines near-infrared (NIR) thermographic sensing with deeplearning algorithms for real-time and accurate detection of subcutaneous veins.Yet another objective of the invention is to assist medical professionals inclinical procedures such as intravenous (IV) catheter placement, blood sampling, andother vascular access techniques by providing enhanced visualization of veinstructures beneath the skin.Further objective of the invention is to enable early diagnosis of vascularabnormalities through AI-powered analysis of thermographic data, thereby improvingprocedural outcomes, patient comfort, and clinical decision-making in varioushealthcare settings.SUMMARY OF THE INVENTIONAccordingly the following invention provides an integrated system for non-invasive subcutaneous vein detection and vascular abnormality diagnosis byintegrating near-infrared (NIR) thermographic imaging with deep learning-basedimage analysis. the system a NIR thermographic sensor array (102), and a high-resolution NIR imaging camera (103) positioned above the patient's hand or forearm(101). The imaging system captures heat-based vascular patterns beneath the skin andtransmits them to the image acquisition and preprocessing module (104), where theraw data is enhanced and normalized for further analysis. This approach enablessuperior vein visibility compared to conventional reflection-based imaging.The enhanced images are processed by a deep learning module (105), whichutilizes convolutional neural networks (CNNs) and segmentation models, includingDeepLabV3+, to identify and map subcutaneous veins and detect vascular anomalieswith high precision. The final annotated output is displayed in real-time on a visualinterface (106), providing medical practitioners with immediate feedback duringprocedures such as IV catheter placement, blood sampling, or preliminary vasculardiagnostics. The system also incorporates adaptive learning capabilities that improvediagnostic accuracy over time through clinical feedback. The integration of thermalimaging and AI-driven segmentation into a compact, user-friendly platform representsa significant advancement in biomedical imaging and patient care.BRIEF DESCRIPTION OF DRAWINGThis invention is described by way of example with reference to the followingdrawings where,Figure 1 of sheet 1 shows the schematic architecture of the integrated NIRthermographic and deep learning system.Where,101 denotes a patient hand,102 denotes a near-infrared thermographic sensors,103 denotes a near-infrared (NIR) camera positioned above the imagingfield,104 denotes an image acquisition and processing module,105 denotes a deep learning processing unit performing segmentation anddisease detection,106 denotes a display screen showing real-time vein mapping anddiagnostic outputs.DETAILED DESCRIPTION OF THE INVENTIONAs used in the description herein and throughout the claims that follow, themeaning of "a," "an," and "the" includes plural reference unless the context clearlydictates otherwise. Also, as used in the description herein, the meaning of "in"includes "in" and "on" unless the context clearly dictates otherwise.The present invention is related to an integrated system for non-invasivesubcutaneous vein detection and vascular abnormality diagnosis. The inventionutilizes a near-infrared thermographic imaging device to capture temperaturevariations and blood vessel profiles beneath the skin. These thermal images areprocessed using advanced deep learning algorithms primarily convolutional neuralnetworks (CNNs) and segmentation models like DeepLabV3+ to accurately highlightsubcutaneous veins and detect pathological conditions. The integrated system outputsreal-time visual overlays to assist healthcare practitioners in clinical interventions andpreliminary diagnostics.The system includes:Near-Infrared Thermographic Sensor Array (102): Emits NIR wavelengths(preferably between 740 nm and 950 nm) to penetrate soft tissue layers and detectheat patterns associated with vascular structures.High-Resolution NIR Imaging Camera (103): Captures reflected and emittedinfrared signals to form a detailed thermographic image of subcutaneous tissues.Image Acquisition and preprocessing unit (104): Processes raw thermographicimages, performing contrast enhancement, noise filtering, and preparation formachine learning input.Deep learning processing Module (105): Applies pre-trained CNN models todetect and segment veins. Utilizes DeepLabV3+ architecture for fine-levelsegmentation and anomaly detection for early disease diagnosis.User Interface and Display Unit (106): Displays annotated images highlightingvein networks and potential vascular anomalies for clinical procedures such as IVplacement, phlebotomy, and preliminary diagnostics.Working of the present embodiment:The device is configured to be positioned at a distance of approximately 15-25centimeters above the patient's hand or forearm (101). Upon activation, the near-infrared (NIR) thermographic sensors capture real-time thermal and vascular imagingdata. The captured images are subsequently preprocessed to normalize brightness,contrast, and enhance visual clarity.The preprocessed images are then input into a deep learning modelspecifically trained to detect subcutaneous vein structures and identify vascularanomalies. The resulting output, including segmented vein maps and identifiedabnormalities, is displayed on a high-contrast display screen, facilitating clinicalprocedures such as intravenous catheterization or serving as preliminary diagnosticinput. The system optionally includes automated alerts to notify clinicians of potentialvascular abnormalities.The invention provides an integrated system for non-invasive subcutaneousvein detection and vascular abnormality diagnosis, configured to operate in real-timeand enhance clinical interventions. The system comprises a near-infraredthermographic sensor array (102) configured to emit NIR wavelengths in the range of740 nm to 950 nm. These wavelengths are capable of penetrating soft tissues to detectthermal signatures associated with subcutaneous vascular structures. A high-resolution near-infrared imaging camera (103) is positioned above the patient hand(101) at a working distance of approximately 15 to 25 centimeters and is configuredto capture reflected and emitted infrared signals to generate detailed thermographicimages of subdermal tissues. The captured images are transmitted to an imageacquisition and preprocessing module (104), which is connected to the imagingcamera (103) and configured to perform contrast enhancement, brightnessnormalization, and noise filtering to optimize image quality.The preprocessed images are then forwarded to a deep learning processingunit (105), connected to the preprocessing module (104), and configured to analyzethe images using a convolutional neural network (CNN) model trained on veinsegmentation and vascular abnormality detection. This unit further comprises asegmentation architecture based on DeepLabV3+, which performs fine-grainedsegmentation of subcutaneous veins and identifies potential vascular anomalies byanalyzing thermal and spatial patterns. A display unit (106), connected to the deeplearning processing unit (105), presents real-time annotated outputs with color-codedoverlays indicating segmented veins and suspected abnormalities. The system isfurther configured to issue automated alerts upon detection of vascular anomalies.The deep learning processing unit (105) incorporates an adaptive learningmechanism that continuously refines CNN model parameters using real-worlddiagnostic feedback. The entire system is implemented as a compact, mobile, andstandalone medical device suitable for bedside use in clinical, surgical, andemergency care settings.Inventive step of your invention: The present invention introduces a novelintegration of near-infrared (NIR) thermographic imaging with AI-powered deeplearning segmentation to enable real-time detection of subcutaneous veins. Unlikeconventional reflection-only NIR imaging systems, this invention utilizes thermaldata, which significantly enhances contrast and improves the visibility of vascularstructures beneath the skin. The system uniquely combines both procedural assistancesuch as intravenous (IV) access and blood sampling with early-stage diagnosis ofvascular abnormalities, all within a single compact, non-invasive platform. A keyinventive feature lies in the implementation of a dual-stage deep learning framework:an initial coarse localization of vein patterns is performed, followed by fine-grainedsegmentation using an advanced deep learning model, specifically the DeepLabV3+architecture. This dual-stage process ensures both high-speed performance anddetailed accuracy in vein mapping and abnormality detection, offering a significanttechnological advancement in biomedical imaging and diagnostic procedures.Industrial application: The invention is useful in healthcare industries such asenhancing clinical efficiency, patient safety and vascular abnormality precision.While various embodiments of the present disclosure have been illustrated anddescribed herein, it will be clear that the disclosure is not limited to theseembodiments only. Numerous modifications, changes, variations, substitutions, andequivalents will be apparent to those skilled in the art, without departing from thespirit and scope of the disclosure, as described in the claims.
Claims
1. An integrated system for non-invasive subcutaneous vein detection and vascular abnormality diagnosis, comprising of; a near-infrared thermographic sensor array (102), configured to emit NIR wavelengths ranging from 740 nm to 950 nm, said wavelengths being capable of penetrating soft tissues to detect thermal signatures associated with subcutaneous vascular structures; a high-resolution near-infrared imaging camera (103), positioned above patient hand (101) at an operational distance of approximately 15 to 25 centimeters, and configured to capture reflected and emitted infrared signals to generate detailed thermographic images of subdermal tissues; an image acquisition and preprocessing module (104), connected to said imaging camera (103), and configured to receive thermographic image data and perform contrast enhancement, brightness normalization, and noise filtering to optimize image quality for further processing; a deep learning processing unit (105), connected to said preprocessing module (104), and configured to process the preprocessed thermographic images using a convolutional neural network (CNN) model trained on vein segmentation and vascular abnormality detection and adaptive learning mechanism that updates the CNN model parameters over time using real-world diagnostic feedback to enhance accuracy and performance; said deep learning processing unit (105) consist of segmentation architecture based on DeepLabV3+, configured to perform fine-grained segmentation of subcutaneous veins and identify potential vascular anomalies by analyzing thermal and spatial image patterns; and a display unit (106), connected to said deep learning processing unit (105), and configured to present real-time annotated visual outputs comprising color-coded overlays representing segmented veins and suspected vascular abnormalities on the patient's thermographic image; wherein said system is further configured to provide automated alerts upon detection of vascular abnormalities.
