IoT Device Identification via Visual Recognition and Neural Networks

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Users of home networks face complexity and difficulty in recognizing and managing network-enabled devices, particularly with the increasing number of IoT devices, due to the lack of meaningful device identification and the need for additional infrastructure, leading to unintentional network administration roles.

Innovation Solution

A method using a mobile communication terminal, such as a smartphone, for object recognition based on image recording and parallel network scanning to identify network-enabled devices, with a central server employing an artificial neural network for device classification and displaying information in augmented reality.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If users manually manage and recognize network devices using traditional methods (lists, browser menus), then device identification information is available, but user complexity and difficulty in recognizing devices increases

Engineering Contradiction:
Improvedevice identification informationVSAvoiduser complexity
Core Design Contradiction:
Loss of informationVSEase of operation

Solution Approach 1:

The patent replaces manual mechanical operations (scrolling through lists, checking browser menus) with automated optical recognition. The mobile terminal's camera captures device images, and an artificial neural network automatically identifies devices, substituting user manual navigation with automated visual recognition processing.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent introduces a mobile terminal as an intermediary between the user and network devices. Instead of users directly interacting with complex network device lists, the mobile terminal serves as a mediator that captures images, processes them through neural networks, and presents simplified identification information to users.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Area of stationary object

If additional infrastructure devices (WLAN repeaters, power lines) are added to extend network coverage, then network coverage is improved, but device complexity and susceptibility to errors increases

Engineering Contradiction:
Improvenetwork coverageVSAvoidinfrastructure complexity
Core Design Contradiction:
Area of stationary objectVSDevice complexity

Solution Approach 1:

The patent extracts the device identification function from the complex network infrastructure and relocates it to a mobile terminal. By taking out the recognition capability from the network infrastructure itself and placing it in the user's mobile device, the system reduces infrastructure complexity while maintaining identification functionality.

Inventive Principle:
Principle #2Taking out (Extraction)

3Quantity of substance

If traditional network device listing methods are used, then all devices can be displayed, but meaningful identification and user understanding of devices becomes difficult

Engineering Contradiction:
Improvenumber of devices displayedVSAvoidmeaningful device identification
Core Design Contradiction:
Quantity of substanceVSLoss of information

Solution Approach 1:

The patent employs visual recognition where devices are identified through their physical appearance, color, shape, and visual characteristics captured by the camera. The neural network analyzes these visual features to provide meaningful identification, transforming abstract device names into visually recognizable entities with distinctive physical attributes.

Inventive Principle:
Principle #32Color changes

Data Source

PatentEP3937430B1Method and system for detecting network devices in a home network
Publication Date: 2022.12.07 DEUTSCHE TELEKOM AG
  • EP3937430B1 patent drawingFigure 1
  • EP3937430B1 patent drawingFigure 2~3
  • EP3937430B1 patent drawingFigure 4

AI summary

To better and/or more easily identify network-enabled devices in a home network and to make information about these devices more and/or more easily accessible to a user, the invention provides a method for identifying a network-enabled device (201-206) in a home network (410), wherein a digital image is captured with a mobile communication terminal (100) and transmitted to a central server (300), in which the image is processed by means of an artificial neural network in order to assign the image to a device class and to transmit corresponding first device class information back to the mobile communication terminal (100).The mobile communication device performs a network scan to discover network-enabled devices on the home network. Depending on the received first device class information, one of the discovered devices is selected, or the user provides second device class information to transmit corresponding second device class information to the central server (300), which is used to train the artificial neural network. The invention further provides a mobile communication device and a system for carrying out the method.