KR-20260062452-A - System for classifying photon limited images to protect personal privacy and method thereof

KR20260062452AKR 20260062452 AKR20260062452 AKR 20260062452AKR-20260062452-A

Abstract

The present invention relates to a photon-limited image classification system and a method for protecting personal privacy. According to the present invention, a photon-limited image classification system for protecting personal privacy comprises: an image processing unit that generates a photon-limited image using a pre-set method on an original image stored in a cloud or on a personal device; and a classification unit that classifies an image by applying a learned classification model to the generated photon-limited image on the personal device, and further comprises a learning unit that trains a classification model through federated learning to classify images for a plurality of photon-limited images, wherein the original image for pre-training the classification model is stored in the cloud, and the original image to be classified by a user may be stored in the personal device. As such, according to the present invention, access to the original image by an attacker or a cloud service provider can be prevented and storage space can be saved.

Inventors

문인규
정온지

Assignees

재단법인대구경북과학기술원

Dates

Publication Date: 20260507
Application Date: 20241029

Claims (10)

An image processing unit that generates a photon-limited image using a pre-configured method on an original image stored in the cloud or on a personal device; and It includes a classification unit that classifies images by applying a learned classification model to the photon-limited image of the generated personal device, and Original images for pre-training the classification model are stored in the above cloud, and The above personal device is a photon-limited image classification system for personal privacy protection in which the original image to be classified by the user is stored.
In paragraph 1, A photon-restricted image classification system for protecting personal privacy, further comprising a learning unit that trains a classification model through federated learning to classify images for multiple photon-restricted images.
In paragraph 1, The above image processing unit is, A photon-limited image classification system for personal privacy protection that generates a photon-limited image by applying Poisson multinomial distribution-based photon counting imaging (PMD-PCI) to an original image stored in the cloud or personal device.
In paragraph 1, The above classification unit is, A photon restriction image classification system for personal privacy protection that classifies images according to the output class by applying a learned classification model to the photon restriction image of the generated personal device.
In paragraph 2, The above learning unit is, A photon-limited image classification system for protecting personal privacy that uses initial weights transmitted from a central server, repeatedly trains a classification model for a number of local epochs, updates the weights of the classification model using the learning rate and loss of the local epochs, and transmits the trained weights of the classification model to the central server to perform weight updates.
A step in which an image processing unit generates a photon-limited image using a pre-configured method on an original image stored in a cloud or personal device; and The classification unit includes the step of classifying an image by applying a learned classification model to the photon-limited image of the personal device in which the classification unit was generated. Original images for pre-training the classification model are stored in the above cloud, and A photon-limited image classification method for protecting personal privacy, wherein the original image to be classified by the user is stored in the above-mentioned personal device.
In paragraph 6, A photon-restricted image classification method for protecting personal privacy, which further includes a step of training a classification model through federated learning so that a learning unit classifies images for multiple photon-restricted images.
In paragraph 6, The step of generating the above is, A photon-limited image classification method for protecting personal privacy, which generates a photon-limited image by applying Poisson multinomial distribution-based photon counting imaging (PMD-PCI) to an original image stored in the cloud or personal device.
In paragraph 6, The above classification step is, A photon restriction image classification method for protecting personal privacy, which classifies images according to the output class by applying a learned classification model to the photon restriction image of the generated personal device.
In Paragraph 7, The above-mentioned learning step is, A photon-limited image classification method for protecting personal privacy, which uses initial weights transmitted from a central server, repeatedly trains a classification model for a number of local epochs, updates the weights of the classification model using the learning rate and loss of the local epochs, and transmits the trained weights of the classification model to the central server to perform weight updates.

Description

System for classifying photon limited images to protect personal privacy and method thereof The present invention relates to a photon-limited image classification system and method for protecting personal privacy, and more specifically, to a photon-limited image classification system and method for protecting personal privacy that generates photon-limited images using Poisson multinomial distribution-based photon counting imaging (PMD-PCI) and classifies the generated photon-limited images through a deep learning model. Recently, the importance of data is increasing as vast amounts of data collected using multiple digital devices are being utilized across various industries. For example, cloud computing platforms such as Google, Amazon, and Microsoft provide scalable storage space and computing power to those who wish to analyze data but lack the financial means to purchase infrastructure. However, there are limitations to privacy protection and security because data outsourced to cloud servers can be stolen by attackers or accessed by cloud service providers. In response to this, although technologies using encryption algorithms to transform data into various forms have been researched to protect data outsourced to cloud servers, cloud service providers and attackers can access the original data because it must be decrypted back to the original before analysis is performed. Therefore, there is a need for technology to generate and classify photon-limited images using a portion of the original data to protect personal privacy within the data. The technology forming the background of the present invention is described in Korean Registered Patent No. 10-2535810 (published May 23, 2023). FIGS. 1 and FIGS. 2 are configuration diagrams of a photon-limited image classification system for personal privacy protection according to an embodiment of the present invention. FIG. 3 is a flowchart of a photon-limited image classification method for protecting personal privacy according to another embodiment of the present invention. FIG. 4 is a diagram briefly illustrating a photon-limited image classification method for protecting personal privacy according to another embodiment of the present invention. FIG. 5 is a diagram illustrating the structure of a classification model according to another embodiment of the present invention. FIG. 6 is a drawing illustrating an example of a photon-limited image according to the original image and the expected number of incident photons according to another embodiment of the present invention. FIG. 7 is a diagram illustrating the classification accuracy of a classification model learned according to a contrast model and another embodiment of the present invention. FIG. 8 is a diagram illustrating the compression performance according to the expected number of incident photons of an original image and a photon-limited image generated according to another embodiment of the present invention. Preferred embodiments according to the present invention will be described in detail below with reference to the attached drawings. In this process, the thickness of lines or the size of components shown in the drawings may be exaggerated for clarity and convenience of explanation. Furthermore, the terms described below are defined in consideration of their functions within the present invention, and these may vary depending on the intent or practice of the user or operator. Therefore, the definitions of these terms should be based on the content throughout this specification. In the embodiment described below, the photon-limited image classification system (100) for protecting personal privacy generates a photon-limited image for an image stored in a personal device (3) and classifies it through a classification model transmitted from a central server (2). The photon-limited image classification system (100) for protecting personal privacy is described by a specific example as being performed by a computing device comprising one or more processors and one or more memories capable of performing the above process. However, the present invention is not limited thereto and may be applied to hosting such as servers and web hosting, and may be applied to data (e.g., documents, videos, etc.) containing personal privacy (e.g., the original author's personal information, etc.) stored in the hosting. FIGS. 1 and FIGS. 2 are configuration diagrams of a photon-limited image classification system for personal privacy protection according to an embodiment of the present invention. As illustrated in FIGS. 1 and 2, a photon-limited image classification system (100) for protecting personal privacy may include an image processing unit (110), a learning unit (120), and a classification unit (130). First, the image processing unit (110) can generate a photon-limited image using a pre-set method (e.g., Poisson multinomial distribution-based photon counting imaging, PMD-PCI) on an original image stored in the cloud (