Description:FORM 2
THE PATENTS ACT 1970
(39 of 1970)
&
The Patents Rules, 2003
COMPLETE SPECIFICATION
(See section 10 and rule 13)

1. TITLE OF THE INVENTION: “A SYSTEM AND METHOD OF LIVE ROLLING FINGERPRINT IMAGE STITCHING WITH STORAGE MEDIUM”
2. APPLICANT:

(a) NAME : Mantra Softech (India) Private Limited
(b) NATIONALITY : Indian
(c) ADDRESS : B-203, 2nd floor, Shapath Hexa
Opp. Gujarat High Court
S. G. Highway, Sola
Ahmedabad 380 060
3. PREAMBLE TO THE DESCRIPTION
PROVISIONAL

The following specification describes the invention. þ COMPLETE

The following specification particularly describes the invention and the manner in which it is to be performed.

Field of the invention
The present invention relates to a system and method of live rolling fingerprint image stitching with storage medium. More particularly, the present invention relates to the technical fields of image processing and pattern recognition, and specifically to a method for real-time stitching of the rolling fingerprint images along with storage medium. The present invention stitches the fingerprint images of each frame based on overlapping fingerprint areas between the fingerprint images.

Background of the invention
Fingerprint recognition technology is currently the most widely used biometric recognition technology. This technology realizes fingerprint recognition based on various feature points contained in the fingerprint, such as end points, bifurcation points, divergence points, isolated points, ring points, short lines, etc. Currently, some high-end mobile phones, tablet computers and other electronic devices have integrated fingerprint recognition functions. Fingerprint recognition technology includes two stages: fingerprint collection (referring to obtaining fingerprint images and extracting fingerprint features, including the collection of fingerprint templates during registration and the subsequent collection of fingerprints to be identified), and fingerprint matching and identification. The existing technology generally collects fingerprints based on press-type fingerprint sensors.
In continuation, press-type fingerprint collection is prone to invalid fingerprint entry problems, such as no overlap between fingerprint areas pressed multiple times, that is, isolated fingerprints, or the overlapping area is too large. So, it is difficult to splice a complete fingerprint, resulting in the user having to press the sensor multiple times to collect fingerprint information that meets normal needs. For example, for a sensor with a larger sensing area (the larger the sensor area, the more fingerprint information can be obtained with a single press), a fingerprint collection generally requires the user to press 5-6 times, while for a sensor with a smaller sensing area, more pressing operations are required. For example, sometimes the user may need to press up to 20 times to obtain complete fingerprint information.
Additionally, with the development of computer technology, biometrics has become the preferred method of identity authentication, among which fingerprint authentication is the most common. The collection area of fingerprint collectors on the market is usually relatively small, and most of them are collected by a flat press method. It is difficult for collectors to collect the same fingerprint area every time, thus increasing the rejection rate of fingerprint identity authentication. Furthermore, fingerprints have uniqueness and permanence because fingerprints are different among individuals and do not change in their lifetime. Thus, fingerprints have been widely used in the situations where identification of an individual is required.
The existing technical problem is when a fingerprint is captured the quality of a captured fingerprint image may vary due to a condition or the like at the capturing and thus a low quality fingerprint image may be obtained. Since a low quality fingerprint image makes identification of an individual difficult, it is desirable for a fingerprint image to be of a higher quality as much as possible. An example of a low quality fingerprint image may be an image with an insufficient contrast between a ridge line part and a valley line part of a fingerprint or an image with uneven brightness. Further, an example of a low quality fingerprint image may be an image in which a part of a fingerprint may be deleted as a result of misalignment of a finger with respect to a region that can be captured by an image sensor at the capturing.
Overall, capturing rolled fingerprints using a fingerprint scanner coupled to a computer may be accomplished in a number of ways. Many current technologies implement a guide to assist the user. This guide primarily discloses two varieties. The first type includes a guide located on the fingerprint scanner itself. This type may include guides such as light emitting diodes (LEDs) that move across the top and/or bottom of the scanner. The user is instructed to roll the finger at the same speed as the LEDs moving across the scanner. In doing so, the user inevitably goes too fast or too slow, resulting in poor quality images. The second type includes a guide located on a computer screen. Again, the user must match the speed of the guide, with the accompanying disadvantages.
Although, conventional efforts to knit image portions into composite fingerprint images typically result in image discontinuities, particularly where image portions overlap and provide different pixel values for overlapping areas. The discontinuities appear particularly at points where ridge features meet in adjacent image portions. There is a need for an improved method to enhance fingerprint recognition systems by stitching together multiple partial fingerprint images captured during a rolling motion.
The present invention has been made in view of the above technical problem and intends to provide a system and method of live rolling fingerprint image stitching with storage medium that can acquire a high quality fingerprint image. In scenarios where capturing the entire fingerprint in a single scan is challenging, stitching techniques are utilized. This method involves capturing multiple partial images as the finger is rolled across the scanner. The system then aligns and merges these images to reconstruct a complete fingerprint and stored in to storage medium. Said approach is particularly applicable for ensuring that all ridge details are captured accurately, even if the initial scans are not comprehensive.
Furthermore, successful stitching for forming the complete fingerprint requires extensive computing resources and the collection of a large plurality of overlapping images. Powerful microprocessors, significant amounts of memory, and a relatively long processing time are required to adequately process the fingerprints.
In order to solve the above technical problems, the present invention discloses the technical solution by providing two models i.e. comprehensive stitching model and a real-time stitching model. The comprehensive stitching model focuses on refining and enhancing the final fingerprint image to ensure its accuracy and quality, making it suitable for secure identification and authentication purposes and the real-time stitching model, on the other hand, is responsible for the immediate processing of fingerprint images as they are captured. An advance machine learning model is used to align overlapping regions by detecting minutiae points and ridge patterns. According to the present invention, through a penalty maps and image stitching model achieve more accurate and visually appealing results, particularly in complex scenes where traditional methods may falter. Additionally, by performing post processing model on the stitched image to enhance clarity and detail required for accurate fingerprint recognition. Said complete fingerprint images are stored in to the storage medium for post processing. It ensures a continuous and seamless user experience during the fingerprint scanning process.
At initial stage of the present invention, as the finger rolls across scanner collect fingerprint images and stitch the images of multiple frames with overlapping areas to obtain a complete and seamless fingerprint image, which can overcome the above defects and increase the effective area of the fingerprint area to provide more fingerprint feature information, thereby make identity authentication more accurate. This present invention technique addresses the limitations of traditional fingerprint capture methods, which often miss parts of the fingerprint, resulting in incomplete and less reliable data.
The present invention method significantly improves the accuracy and reliability of fingerprint-based security systems used in various applications, including mobile devices, access control, and financial transactions. Law enforcement agencies can benefit from this technique by reconstructing high-quality fingerprint images from partial prints found at crime scenes, aiding in criminal investigations. Additionally, it enhances the accuracy of identity verification systems used in high-security environments like airports and border control, ensuring that individuals are correctly identified. By capturing multiple partial images and seamlessly stitching them into a high-resolution fingerprint, this innovative approach provides a robust solution for enhancing fingerprint recognition systems, making them more accurate and reliable for a wide range of applications in security, identification, and forensic analysis.

Object of the invention
The main object of the present invention is to provide a system and method of live rolling fingerprint image stitching with storage medium.
Another object of the present invention is to provide a live rolling fingerprint stitching method, which can realize live-time stitching, improve the fingerprint stitching effect and generate a more complete fingerprint, thereby improving identification success rate of the fingerprint collection device.
The other object of the present invention is to perform advanced check to detect and correct horizontal and vertical slippage, to verify image continuity to ensure correct alignment of successive frames.
Still other object of the present invention is to provide live preview by downsizing selected frames to identify an optimal stitching path using machine learning model.
The other object of the present invention is stitching the multiple frames of images collected during finger scrolling to obtain a complete fingerprint image and provide more fingerprint information.
Further object of the present invention is to provide live-time processing, the entire method, from image capture to verification is suitable for applications requiring quick and reliable fingerprint recognition, such as security systems and biometric authentication.
The further object of the present invention is to improve the accuracy and reliability of fingerprint-based security systems used in various applications, including mobile devices, access control, and financial transactions.
Another object of the present invention is to provide comprehensive stitching model and live-time stitching model along with the storage medium to store the final fingerprint image.
The other object of the present invention is to provide the comprehensive stitching method for fingerprint images aim to create high-quality, unified representations by integrating multiple partial scans. Said methods address challenges such as skin elasticity, image quality variations, and the self-similar nature of fingerprint ridge patterns.
The other object of the present invention is to provide the overlapping areas to create smooth transitions, resulting in continuous stitches the aligned images together by blending and complete fingerprint image through machine learning model.
Yet another object of the present invention is to provide a robust solution for enhancing security and efficiency in biometric identification process.
Another object of the present invention is to provide a powerful and reliable method for secure identify verification in diverse applications.
Still another object of the present invention is to provide multi-faceted approach by combining live time stitching and comprehensive stitching method which more reliably and accurately provide final fingerprint image.

Summary of the Invention
The present invention relates to a system and method of live rolling fingerprint image stitching with storage medium. The present invention comprises rolling fingerprint scanner, rolling fingerprint stitching module, a display module and a storage medium. The rolling fingerprint scanner captures multiple overlapping partial fingerprint images as a finger moves over scanner and aligns these overlapping regions by detecting minutiae points and ridge patterns through machine learning model. The rolling fingerprint stitching module stitches capture rolls frames together by blending overlapping areas to create a continuous and complete fingerprint image. Said complete fingerprint images are stored in to storage medium and displays real-time fingerprint image via display module. Said modules are interfaced with rolling fingerprint stitching module. Said module comprises comprehensive stitching method focuses on refining and enhancing final fingerprint image to ensure its accuracy and quality, making it suitable for secure identification and authentication purposes and real-time stitching module responsible for immediate processing of fingerprint images as they are captured and ensures continuous, seamless user experience during fingerprint scanning process. The final images undergo verification against database of known fingerprints, confirming parameters. Perform post-processing to enhance clarity and accurate fingerprint recognition via spatial and frequency domain filtering model.

Brief Description of the Drawings
FIG. 1 is a block diagram of a fingerprint scanner hardware configuration according to present invention.
FIG. 2 illustrates a flowchart of comprehensive stitching model along with storage medium according to the present invention.
Fig. 3 illustrates the flowchart of real-time stitching model with storage medium according to the present invention.
Fig. 4 illustrates a flow chart of performing post-processing according to the present invention.

Detailed description of the Invention
Before explaining the present invention in detail, it is to be understood that the invention is not limited in its application. The nature of invention and the manner in which it is performed is clearly described in the specification. The invention has various components and they are clearly described in the following pages of the complete specification. It is to be understood that the phraseology and terminology employed herein is for the purpose of description and not of limitation.
As used herein, the term "module", and “model” refers to as the unique and addressable components of the software implemented in hardware which can be solved and modified independently without disturbing (or affecting in very small amount) other modules of the software implemented in hardware.
As used herein, the term "database" refers to either a body of data, a relational database management system (RDBMS), or to both. The database includes any collection of data including hierarchical databases, relational databases, flat file databases, object-relational databases, object oriented databases, and any other structured collection of records or data that is stored in a computer system/storage medium. The above examples are example only, and thus are not intended to limit in any way the definition and/or meaning of the term database.
As used herein, the term “device” or “scanner”, refers to a unit of hardware, outside or inside the case or housing that is capable of providing input or of receiving output or of both. It also includes a fingerprint senssor such as smart phone, tablet, laptop etc.
As used herein, the term “live rolling fingerprint image stitching”, refers to a real-time image stitiching per se incorporated herein the system as well as method disclosure of the present invention.
Now, FIG. 1 describes a block diagram of a rolled fingerprint scanner hardware configuration according to present invention. A system and method of live rolling fingerprint image stitching with storage medium includes rolling fingerprint scanner, rolling fingerprint stitching module, a display module and a storage medium. The rolling fingerprint scanner captures full user fingerprint image by rolling the finger across a flat surface. Said rolled fingerprint scanner detects start and stop of fingerprint rolls on fingerprint scanner, as the finger rolls across the scanner, it captures a plurality/series of overlapping partial fingerprint images. Simultaneously, the rolling fingerprint stitching module combine/stitches the capture rolls images/frames together by blending the overlapping areas to create a continuous and complete fingerprint image. Said complete fingerprint images are stored in to the storage medium and displays live fingerprint image via display module. The storage medium and the display module are interfaced with the rolling fingerprint stitching module. Simultaneously, perform initial assessment on the captured images/frames to detect blank frames, frames containing multiple fingerprints and blurriness. Further, selecting frames that meet predetermined quality thresholds and stored frames in a storage medium for post processing and generating a live preview via display module by downsizing selected frames.
The next step of the present invention is calculating penalty map on the downsized images to identify an optimal/shortest stitching path through Dijkstra model and display live preview of line-based stitching output with a colored vertical indicator reflecting centroid and stitch quality An image quality detection module integrates to evaluate each captured image based on quality metrics and generate an image quality score for each image. An image correction module dynamically corrects images that do not meet predetermined quality thresholds and establish self adaptive feedback loop model continuously adjust both the image capture parameters and processing in real-time based on quality assessments. Stitches the quality frames together to form a complete fingerprint image and perform post-processing on the stitched image through both conventional spatial and frequency domain filtering model in combination with the image correction model to generate a final high-resolution fingerprint image.
The rolling fingerprint stitching module comprises comprehensive stitching method and real-time stitching method. As shown in FIG. 2 of the present invention, the comprehensive stitching method comprises:
S11: capturing plurality of overlapping partial fingerprint frames by rolling fingerprint scanner;
S12: reading pairs of fingerprint frames along with their foreground mask pairs;
In step S12, the foreground mask pairs determine a dataset containing pairs of original images and their corresponding foreground masks. Said foreground mask means, a mask separates the foreground from the background. A mask is a technique used to isolate certain parts of an image while hiding or ignoring others.
S13: performing initial assessment on the captured images/frames to detect blank frames, frames containing multiple fingerprints and blurriness;
S14: evaluating rolling speed and rolling direction of the fingerprint by analyzing the overlapping areas of adjacent frames through machine learning model;
In step S13 and S14, performs several preliminary checks on the frames. Said checks determining if any frames are blank, if there are multiple fingerprints present in the same frame, or if the frames are blurry. Additionally, the machine learning model assesses the rolling direction of the fingerprint and evaluates the rolling speed by examining the overlapping areas of adjacent frames. These checks are crucial for ensuring that only high quality frames are processed further. After getting the high quality frames, it normalizes contrast of successive/second image to match the first image by adjusting mean and variance to ensure a consistent visual appearance of the final stitched image.
S15: executing advanced checks to detect and correct horizontal and vertical slippage to verify image continuity through machine learning model;
In S15, it continuously monitors changes in the fingerprint pattern during capture to detect and correct slippage, both horizontally and vertically to ensure that the capture frame is nested within the previous image frame. Thereby ensuring correct alignment of successive frames and maintaining the continuity of the fingerprint image.
S16: selecting frames that pass/meet predetermined overlap criteria and stored frames in a storage medium for post processing;
Furthermore, in S16, selected frames that pass predetermined checks, stored in a common vector with an approximate 70% overlap. Said overlap is essential for creating a seamless final image. In next step, for live preview purposes, said frames are downsized by a factor of 2, allowing the user to see a real-time representation of the stitching process. Further, the downsizing factor for the live review is adjustable to balance processing speed with preview clarity. The downsizing means an image by a factor of 2 reduces its dimensions by half, resulting in a smaller file size and lower resolution. This process involves averaging groups of four neighboring pixels to create a single pixel in the resized image, effectively reducing both width and height by half.
S17: generating a live preview by downsizing the selected frames;
S18: calculating penalty map on the downsized first image and the contrast-matched second image to identify a shortest stitching path;
In above mentioned step S18, a penalty map is a calculated representation of the downsized images, used to identify optimal paths for seamlessly merging frames. This technique is particularly valuable in regions with complex textures or significant depth variations, where traditional stitching methods may struggle. Said penalty map calculation includes weighting factors based on ridge and valley patterns of the fingerprint images to enhance the accuracy of the stitching path determined by the Dijkstra model.
S19: stitching the images together through Dijkstra model determine the shortest and most accurate path.
S20: providing live preview on display, a line-based stitching output with a colored vertical line changes from red to green indicates total stitch width meets /exceeds required threshold for a high quality fingerprint image.
In above Step 19 and step 20, stitch the selected frames along the determined optimal path to generate a continuous fingerprint image. Furthermore, a colored vertical rectangular indicates estimated centroid of the stitched fingerprint and provide real-time feedback on stitch quality. Said colored rectangular box in the live preview changes the color from red to green when the total stitch width meets a predetermined threshold indicative a high quality fingerprint image.
Further in next step S21, perform post processing on the stitched image includes spatial and frequency domain filtering to enhance the quality and clarity of the final fingerprint image. It makes suitable for accurate fingerprint recognition and authentication.
S21: performing post-processing on the final stitched image by spatial and frequency domain filtering model to enhance the clarity for accurate fingerprint recognition.
The main aspect of present invention involves a process for real time stitching method according to embodiment of the present invention as shown in FIG. 3. The process comprises:
S31: capturing plurality of overlapping partial fingerprint images/frames by rolling fingerprint scanner;
S32: reading pairs of fingerprint images/frames along with their foreground mask pairs;
S33: performing initial assessment on the captured frames to detect blank frames, the frames containing multiple fingerprints and blurriness;
S34: evaluating rolling speed and rolling direction of the fingerprint by analyzing the overlapping areas of adjacent frames through machine learning model;
S35: executing advanced checks to detect and correct horizontal and vertical slippage to verify image continuity through machine learning model;
S36: selecting frames that meet predetermined quality thresholds and stored frames in a storage medium for post processing;
S37: generating a live preview by downsizing the selected frames;
S38: calculating penalty map on the downsized first image and the contrast-matched second image to identify a shortest stitching path through Dijkstra model;
S39: providing live preview on display, a line-based stitching output with a colored vertical line changes from red to green indicates total stitch width meets required threshold for a high quality fingerprint image.
In continuation, in next step of the Fig. 3 of the present invention, the real-time stitching method is to ensure the high-quality image. Said process applies enhancement techniques like noise level; contrast adjustment; ridge clarity & uniformity and overall sharpness to generate an image quality score for each image. An image quality detection module is integrated in to the real-time stitching model to evaluate each captured image based on quality metrics.
S40: integrating an image quality detection module to evaluate each captured image based on quality metrics and generate an image quality score for each image;
S41: applying an image correction module to automatically correct frames that do not meet predetermined quality thresholds;
In above step S41, the image correction module is dynamically applied for advanced image processing techniques, including noise reduction, contrast enhancement, super resolution, and artifact removal, to automatically correct images that do not meet predetermined quality thresholds.
S42: establishing self adaptive feedback loop model;
In S42, the machine learning modules continuously adjust both the image capture parameters and the processing model of real time based captured frames on evolving quality assessments and establishes a self adaptive feedback loop. After, receiving the quality enhance image, its post processing for the stitching means a complete fingerprint image. S43: stitching the quality frames together to form a complete fingerprint image;
S44: performing post-processing on the stitched image through both spatial and frequency domain filtering model in combination with image correction model to generate a final high-resolution fingerprint image.
In above mentioned step S44, the post processing includes both the spatial and frequency domain filtering to enhance the clarity of the final stitched fingerprint image.
The next step is shown in fig. 4 of the present invention performing the post processing on the final image. The final images undergo verification against database of known fingerprints, confirming parameters that stored in the storage medium and perform post-processing to enhance clarity and accurate fingerprint recognition via spatial and frequency domain filtering model. The spatial and frequency domain filtering techniques are two fundamental approaches used in image processing for tasks like noise reduction, edge detection, and image enhancement.
The present invention system and method for live rolling fingerprint stitching present an innovative and comprehensive approach to enhancing the accuracy and reliability of fingerprint recognition systems. The present invention method’s, comprising real-time stitching model and comprehensive stitching model, ensures that fingerprint images are processed and enhanced efficiently and effectively. Real-time stitching handles the immediate processing of fingerprint images, performing preliminary and advanced checks, storing selected frames and providing a live preview. The comprehensive stitching refines the final fingerprint image, ensuring consistency and quality through contrast normalisation, the Dijkstra algorithm, and post-processing techniques.
Furthermore, said method's ability to capture multiple partial images and seamlessly stitch them into a high-resolution fingerprint image makes it suitable for various applications in security, identification, and forensic analysis. It is particularly beneficial for biometric authentication in mobile devices, access control, financial transactions, and law enforcement. Law enforcement agencies can reconstruct high-quality fingerprint images from partial prints found at crime scenes, aiding in criminal investigations. Additionally, it enhances the accuracy of identity verification systems used in high security environments like airports and border control, ensuring that individuals are correctly identified.
While various elements of the present invention have been described in detail, it is apparent that modification and adaptation of those elements will occur to those skilled in the art. It is expressly understood, however, that such modifications and adaptations are within the spirit and scope of the present invention as set forth in the following claims.

, Claims:We Claim:
1. A system of live rolling fingerprint image stitching with storage medium comprising:
the rolling fingerprint scanner captures plurality of overlapping partial fingerprint images of user rolling finger across a flat surface;
the rolling fingerprint stitching module stitches the capture rolls frames by blending the overlapping areas to create a continuous and complete fingerprint image;
the initial assessment on captured frames to detect blank frames, the frames containing multiple fingerprints and blurriness;
selecting frames that meet predetermined quality thresholds and stored frames in a storage medium for post processing
generating a live preivew via display module by downsizing selected frames;
characterized in that
calculating penalty map on the downsized images to identify an shortest stitching path through Dijkstra model;
display live preview of line-based stitching output with a colored vertical indicator reflecting centroid and stitch quality;
an image quality detection module is integrate to evaluate each captured image based on quality metrics and generate an image quality score for each image;
an image correction module dynmically apply to automatically correct frames that do not meet predetermined quality thresholds;
establish self adaptive feedback loop model to adjust both the image capture parameters and processing of real-time based captured frames on quality assessments;
stitches the above extracted quality frames together to form a complete fingerprint image;
perform post-processing on the stitched image through both conventional spatial and frequency domain filtering model in combination with the image correction model to generate a final high-resolution fingerprint image.
2. The system of live rolling fingerprint image stitching with storage medium as claimed in claim 1, wherein the storage medium and the display module are interfaced with the rolling fingerprint stitching module.
3. The system of live rolling fingerprint image stitching with storage medium as claimed in claim 1, wherein final image undergo verification against database of known fingerprints stored in to the storage medium.
4. The system of live rolling fingerprint image stitching with storage medium as claimed in claim 1, wherein the image detection module is trained on a dataset of fingerprint images accurately predict quality metrics.
5. The system of live rolling fingerprint image stitching with storage medium as claimed in claim 1, wherein the image correction module dynamically adjusts correction parameters based on real-time quality score to optimize the visual clarity of the fingerprint images.

6. The system of live rolling fingerprint image stitching with storage medium as claimed in claim 1, wherein the self-adaptive feedback loop further modifies the capture settings of the rolling fingerprint scanner to mitigate low-quality image acquisition based on the real-time performance.
7. A method of live rolling fingerprint image stitching with storage medium comprising real-time stitching method steps:
capturing plurality of overlapping partial fingerprint frames by rolling fingerprint scanner;
reading pairs of fingerprint frames along with their foreground mask pairs;
performing initial assessment on the captured frames to detect blank frames, frames containing multiple fingerprints and blurriness;
evaluating rolling speed and rolling direction of the fingerprint by analyzing the overlapping areas of adjacent frames through machine learning model;
executing advanced checks to detect and correct horizontal and vertical slippage to verify image continuity through machine learning model;
selecting frames that meet predetermined quality thresholds and stored frames in a storage medium for post processing;
generating a live preview by downsizing the selected frames;
calculating penalty map on the downsized first image and the contrast-matched second image to identify an optimal/shortest stitching path through Dijkstra model;
wherein,
integrating an image quality detection module to evaluate each captured image based on quality metrics and generate an image quality score for each image;
applying an image correction module to automatically correct images that do not meet predetermined quality thresholds;
establishing self adaptive feedback loop model;
stitching the quality frames together to form a complete fingerprint image;
performing post-processing on the stitched image through both spatial and frequency domain filtering model in combination with image correction model to generate a final high-resolution fingerprint image
8. The method of live rolling fingerprint image stitching with storage medium as claimed in 7, wherein comphrehensive stitching method steps:
capturing plurality of overlapping partial fingerprint frames by rolling fingerprint scanner;
reading pairs of fingerprint frames along with their foreground mask pairs;
performing initial assessment on the captured frames to detect blank frames, frames containing multiple fingerprints and blurriness;
evaluating rolling speed and rolling direction of the fingerprint by analyzing the overlapping areas of adjacent frames through machine learning model;
executing advanced checks to detect and correct horizontal and vertical slippage to verify image continuity through machine learning model;
selecting frames that meet predetermined overlap criteria and stored frames in a storage medium for post processing;
generating a live preview by downsizing the selected frames;
calculating penalty map on the downsized first image and the contrast-matched second image to identify an shortest stitching path;
stitching the images together through Dijkstra model determine the shortest and most accurate path;
providing live preview on display, a line-based stitching output with a colored vertical line changes from red to green indicates total stitch width meets required threshold for a high quality fingerprint image;
performing post-processing on the final stitched image by spatial and frequency domain filtering model to enhance the clarity for accurate fingerprint recognition.
9. The method of live rolling fingerprint image stitching with storage medium as claimed in claim 7, wherein the real-time stitching method determine immediate processing of fingerprint images as captured and ensure a continuous complete fingerprint images.
10. The system of live rolling fingerprint image stitching with storage medium as claimed in claim 8, wherein the comphrehensive stitching method determine refining and enhancing the final fingerprint image to ensure accuracy and quality.
11. The method of live rolling fingerprint image stitching with storage medium as claimed in claim 8, wherein normalising contrast of successive fimngerprint images by adjusting mean and variance to ensure a consistent visual appearnce in the final stitched image.

12. The method of live rolling fingerprint image stitching with storage medium as claimed in claim 8, wherein downsizing determining the image by a factor of 2 reduces its dimension by half for the live preview to balance processing speed with preview clarity.
13. The method of live rolling fingerprint image stitching with storage medium as claimed in claim 8, wherein the penalty map determining calculation representation of downzied images to identify optimal paths for stitching frames.
14. The method of live rolling fingerprint image stitching with storage medium as claimed in claim 8, wherein storing final images in to the storage medium for post processing.

Dated this on 31st day of May 2025