FORM 2
THE PATENTS ACT, 1970
(39 of 1970)
&
THE PATENTS RULES, 2003
COMPLETE SPECIFICATION (See section 10, rule 13)
1. Title of the invention: TREAD DEPTH MEASUREMENT
2. Applicant(s)
NAME NATIONALITY ADDRESS
CEAT LIMITED Indian CEAT Ltd At: Get Muwala Po: Chandrapura Ta: Halol - 389 350 Dist: Panchmahal, Gujarat, India
3. Preamble to the description
COMPLETE SPECIFICATION
The following specification particularly describes the invention and the manner in which it
is to be performed.

FIELD OF INVENTION
[0001] The present invention relates generally to tires and more
particularly to tires having treads with circumferential grooves.
BACKGROUND
[0002] There may be various categories of tires, such as passenger car
tires, truck bus tires, light truck tires, farm use tires, etc. Each of these tires have a tread portion or 'tread' as it is usually called, that is generally made up of several patterned rubber units evenly distributed along a circumference of the tire. It is the tread portion of the tire that makes contact with a road or ground surface during the rolling of the tire. The tread portion of the tire plays an important role in the driving safety of a vehicle into which the tire is fitted. For instance, the tread portion may increase braking force and/or driving force, avoid sliding, or may drain water and keep grip ability, and so on.
[0003] The tread portion of the tires is often molded in such a way as to
create a pattern of grooves that provide means for water evacuation and form the biting edges of tread elements that give the tire traction on the road surfaces. Grooves are elongated void areas in the tread portion that may extend circumferentially or laterally about the tread portion in a straight, curved, or zigzag manner. Grooves are typically wide enough to remain open as the tire rolls through the portion of the tread contacting the road surface.
[0004] Tread depth also called the "non-skid depth" of a tread, is a
distance between an outer surface of the tread portion and deepest grooves as measured at a groove base or bottom. Often these deepest grooves are circumferentially continuous grooves but can be inclined or lateral extending grooves depending on the tread pattern. A shallow tread portion, i.e., a tread portion having inadequate non-skid depth in the grooves of the tread portion may cause traffic hazards and even lead to major accidents. For example, if the tread portion wears beyond a certain depth, the water drainage paths

become inefficient in removing the water due to a decrease in the size of the grooves. This may cause hydroplaning that may have severe consequences. Thus, it is recommended to replace the tire when the tread portion wears beyond a certain depth.
BRIEF DESCRIPTION OF THE DRAWINGS
[0005] The following detailed description references the drawings,
wherein:
[0006] Figure 1 illustrates a network environment implementing a tread
depth measurement system to measure a tread depth of a tire, in
accordance with an example implementation of the present subject matter;
[0007] Figure 2 illustrates the tread depth measurement system coupled
to a training data generation system, in accordance with another example
implementation of the present subject matter;
[0008] Figure 3 illustrates the training data generation system, in
accordance with another example implementation of the present subject
matter;
[0009] Figure 4 illustrates a tire having a plurality of grooves in a tread
portion, in accordance with another example implementation of the present
subject matter; and
[0010] Figure 5 illustrates a tread depth measurement method for
measuring a tread depth of a tire, in accordance with an example
implementation of the present subject matter.
[0011] Throughout the drawings, identical reference numbers designate
similar, but not necessarily identical, elements. The figures are not
necessarily to scale, and the size of some parts may be exaggerated to
more clearly illustrate the example shown. Moreover, the drawings provide
examples and/or implementations consistent with the description; however,
the description is not limited to the examples and/or implementations
provided in the drawings.

DETAILED DESCRIPTION
[0012] As a tire wears out, tread depth of the tire decreases. This may result
in a decrease in some desired tire characteristics, such as wet skid, wet
handling, etc. Hydroplaning characteristics also decrease due to a reduction
in groove volume resulting from the increased wear of tread of the tire. If the
tread profile is too worn or if the tread depth is too low, the tire may lose traction
more easily, leading to accidents involving property and personal injury.
[0013] Therefore, it is desirable to measure the tread depth of the tire to
have an assessment of the tread depth, for example, to provide a user of the tire a recommendation when the tread depth has reached a value that indicates a safe state of a vehicle can no longer be guaranteed for the vehicle in which said tire is fitted.
[0014] Many solutions for measuring the tread depth of a tire are known in
the prior art. For example, one prior art solution for measuring the tread depth of the tire includes using tread depth gauges. A tread depth gauge comprises a probe sliding in a cylinder. The probe is inserted into a groove of the tire and the cylinder advanced till its end is in contact with the tread. The tread depth is read off a scale, which in some instances, is on the other end of the probe projecting from the other end of the cylinder.
[0015] Another prior art solution for measuring the tread depth of the tire
includes observing a tread profile of the tire across one measuring line by a laser or light emanating from a light source in a fan beam pattern, which is reflected to a sensor. The signal of the reflected fan beam is evaluated using a triangulation method to determine a measure of the tread depth of the tire across the illuminated line of the tire surface.
[0016] Yet another prior art solution for measuring the tread depth of the
tire involves making an image from structured light illumination that may be
analyzed to give tire surface contour information from which the tread depth
may be calculated. The image may be taken across the width of the tire to
detect uneven wear.
[0017] Such prior art solutions for measuring the tread depth of the tire are

highly subjective as they either vary in accuracy depending upon the skill of a person tasked with measuring the tread depth or are dependent on the condition of a road surface or the tread depth changes and thus may provide a non-accurate measurement of the tread depth. Further, such prior art solutions for measuring the tread depth of the tire often involve the use of sophisticated tools/mechanical devices that make the process of tread depth measurement cost intensive.
[0018] Thus, there is a need for a technique using which the tread depth
measurement may be carried out easily and objectively so that the measurements and corresponding observations if any, vis-à-vis the tread depth of the tire may be made available to an owner of the vehicle for appropriate action.
[0019] According to various aspects of the present subject matter, systems
and methods for measuring the tread depth of a tire are described. In an example, the systems and methods for measuring the tread depth of the tire the correct tread depth of the tire may be computed, thereby reducing the manual intervention, and increasing the efficiency as well as the accuracy of the tread depth measurement process. The tread depth measurement system may be used for non-invasive and accurate measurement of the tread depth of the tire without requiring any manual intervention or the use of external tools.
[0020] In an example implementation, the tread depth measurement
system comprises a communication module that receives a target image corresponding to a request for calculating the tread depth of a tire. The request for calculating the tread depth of the tire may be raised by an individual, such as the owner or user of a vehicle into which said tire is fitted who wishes to inquire about the tread depth. The tread depth measurement system further comprises an object detection module. The object detection module is configured to analyze the target image to detect if the target image has a tire with a tread portion. The tread portion of the tire may have a plurality of tread elements wherein a shape of these tread elements may be defined by a

plurality of grooves. Once the object detection module confirms the presence of a tire with a tread portion in the target image, the system proceeds to assess the tread depth of the tire in the target image.
[0021] For this purpose, the tread depth measurement system includes a
groove points detection module. The groove points detection module is trained based on machine learning algorithms using historical data collected from a test environment to detect at least one upper groove point and at least one lower groove point of at least one groove of the plurality of grooves present in the tread portion of the tire detected in the target image. The upper groove point corresponds to a position on an outermost surface of the tread portion. Whereas the lower groove point corresponds to a groove bottom of the at least one groove of the plurality of grooves.
[0022] A tread depth determination module of the tread depth
measurement system determines a distance between the at least one upper and lower groove point of the at least one groove of the plurality of grooves, which corresponds to the tread depth of the tire.
[0023] The tread depth measurement system as described herein employs
machine learning models to allow the use of a set of historical data as a training set that yields a particular analytical or statistical model. Applying this analytical or statistical model corresponding to a request for calculating the tread depth of a tire may allow the model to be retrained to provide more accurate tread depth calculation.
[0024] Thus, the present invention enables providing a user with
reasonably accurate measurements of the tread depth of the tire on a real-time basis without requiring the use of any destructive method or external tools. The use of machine learning in the present system also ensures minimal dependency on human expertise, thereby reducing errors in the tread depth measurement as well as the cost of tire warranty claims for the tire manufacturers.
[0025] The above-described tread depth measurement system for
measuring the tread depth of a tire is further described with reference to

Figures 1 to 5. It should be noted that the description and figures merely illustrate the principles of the present subject matter along with examples described herein and should not be construed as a limitation to the present subject matter. It is thus noted that various arrangements may be devised that, although not explicitly described or shown herein, describe the principles of the present subject matter. Moreover, all statements herein reciting principles, aspects, and examples of the present subject matter, as well as specific examples thereof, are intended to encompass equivalents thereof.
[0026] Figure 1 illustrates a network environment 100 comprising a tread
depth measurement system 102 (also referred to as 'system 102') for measuring a tread depth of a tire, according to an example of the present subject matter. The system 102 provides a user, for example, an owner of a vehicle fitted with said tire or a representative of a tire manufacturing company, a platform that utilizes machine learning models to calculate the tread depth of the tire. This non-contact approach may help determine the correct tread depth of the tire and reduce manpower. This approach may also reduce the expense and improve the convenience of measurement of the tread depth of the tire.
[0027] The network environment 100 may comprise a plurality of user
devices 104-1, 104-2, …… and 104-n that are coupled to the system 102, according to an example of the present subject matter. Examples of the user devices 104-1, 104-2, …, 104-n may include but are not limited to, electronic devices, such as a desktop computer, a laptop, a smartphone, a personal digital assistants (PDAs), and a tablet. For example, the user devices 104-1, 104-2, …, 104-n may include an image capturing device, such as a camera or may be coupled to an image capturing device. In an example, the user devices 104-1, 104-2, …, 104-n may communicate with the system 102 over a network 106.
[0028] In an example, the network 106 may be a single network or a
combination of multiple networks and may use a variety of different

communication protocols. The network 106 may be a wireless or a wired
network, or a combination thereof. Examples of such individual networks
include, but are not limited to, Global System for Mobile Communication
(GSM) network, Universal Mobile Telecommunications System (UMTS)
network, Personal Communications Service (PCS) network, Time Division
Multiple Access (TDMA) network, Code Division Multiple Access (CDMA)
network, Next Generation Network (NON), Public Switched Telephone
Network (PSTN). Depending on the technology, the network 106 may
include various network entities, such as gateways, and routers, however,
such details have been omitted for sake of brevity of the present description.
[0029] In an example, the system 102 may be implemented as any of a
variety of conventional computing devices, including, a desktop, a personal computer, a notebook or portable computer, a workstation, a mainframe computer, a cloud environment, and a laptop. Further, in one example, the system 102 may be a distributed or centralized network system in which different computing devices may host one or more of the hardware or software components of the system 102. In another example, the system 102 may host hardware and software components implementing techniques of measuring the tread depth of the tire by processing an image having a tire with a tread portion.
[0030] The network environment 100 may further include a central server
108. The central server 108 may be coupled to an object store 110 configured to store the various information that may be received, exchanged, generated, or stored to compute a measurement of the tread depth of the tire, referred to as tread depth measurement information. In some embodiments, the object store 110 may be internal to the central server 108. In some embodiments, the system 102 may communicate with the object store 110 either directly or over the network 106. In some alternative embodiments, the system 102 may itself include a data store that may serve as a repository to store the tread depth measurement information. In some other alternative embodiments, the system 102 may

also be part of a hosted service executed on the central server 108.
[0031] The system 102 may be configured to receive, process, and
transmit the tread depth measurement information over the network 106 and store the tread depth measurement information in the object store 110 that is coupled to the central server 108. The central server 108 may thus serve as a repository of electronic, computer-readable information for the tire manufacturer, a field representative of the tire manufacturer, a channel partner, and/or a likely user seeking to measure tread depth of a tire that is fitted into a vehicle of the user, as examples. It is to be understood that in the present context, a channel partner may be a person or a company that partners with the tire manufacturer to market and sell the tire manufacturer's products, such as tires.
[0032] In an example, a tread depth measurement client 112 may be
locally provided as a web or mobile application on any of the plurality of user devices 104-1, 104-2, …, 104-n, and used to enable the user to access the system 102 to measure the tread depth of the tire. In an alternative embodiment, any of the plurality of user devices 104-1, 104-2, …, 104-n may access the system 102 using web browsers or mobile apps to measure the tread depth of the tire via a web-based interface provided by the system 102. The user devices 104-1, 104-2, …, 104-n may be configured to receive inputs from users and communicate said inputs to the system 102, or components thereof for processing said inputs for measuring the tread depth of the tire.
[0033] Based on inputs received from the users, such as the owner of the
vehicle, or a representative of the tire manufacturing company, through any of the plurality of user devices 104-1, 104-2, …, 104-n and information available in the object store 110, the process of generating a measurement of the tread depth of a tire may be performed by the system 102 using machine learning model on a real-time basis. Thus, by using the system 102, the tread depth of the tire may be determined with high accuracy and high reliability. For further explanation of the implementation and operation

of the system 102 to compute a measurement of the tread depth of the tire, a reference is made to Figures 2 and 4.
[0034] In an example embodiment, the network environment 100 may
also include a remote access machine (not illustrated), similar to the user devices 104-1, 104-2, …, 104-n. The remote access machine may be configured to provide a user interface to allow authorized individuals, such as channel partners, or organizations to communicate with the central server 108 to access the tread depth measurement information. The remote access machine may include a central server access client (not illustrated), such as a web browser or a web application, that executes on the remote access machine and accesses the central server 108 via a network such as, for example, the network 106.
[0035] Figure 2 illustrates the system 102 coupled to a training data
generation system 232, according to an example implementation of the present subject matter. The training data generation system 232 provides information needed for the system 102 to be able to compute the measurement of the tread depth of the tire when requested by a user of the system 102. Figure 3 illustrates the training data generation system 232 in detail, in accordance with another example implementation of the present subject matter. Figure 4 illustrates a tire 400 having a tread portion 402 in accordance with an implementation of the present subject matter. For sake of ease of understanding of the present invention, Figs. 2 and 4 are explained together.
[0036] The system 102 of the present invention provides a non-contact
approach to calculating the tread depth of a tire, for example, the tire 400 as shown in Figure 4, and also reduces manpower expense and improves the convenience of measurement of the tread depth.
[0037] As depicted in Figure 2, in an example implementation, the system
102 may include at least one processor 202 and a memory 204 coupled to the processor 202. In an example, the processor 202 may be implemented as microprocessors, microcomputers, microcontrollers, digital signal

processors, central processing units, state machines, logic circuitries, and/or any devices that manipulate signals based on operational instructions. The memory 204 may include any computer-readable medium known in the art including, for example, volatile memory (e.g., RAM), and/or non-volatile memory (e.g., EPROM, flash memory, etc.). The memory 204 may also be an external memory unit, such as a flash drive, a compact disk drive, an external hard disk drive, or the like.
[0038] Also, as depicted in Figure 2, in an example implementation,
interface(s) 206 may be coupled to the processor 202. The interface(s) 206 may include a variety of software and hardware interfaces that allow interaction of the system 102 with other communication and computing devices, such as network entities, external repositories, and peripheral devices. The interface(s) 206 may also enable the coupling of components of the system 102 with each other. Further, in an example, the interface(s) 206 may couple the user devices 104-1, 104-2, …, 104-n to the system 102. Likewise, the interface(s) 206 may couple the central server 108 and the system 102. The interface(s) 206 may also enable coupling of internal components of the system 102 with each other.
[0039] The system 102 may also comprise module(s) 208 and data 224
coupled to the processor 202. In one example, the module(s) 208 and data 224 may reside in the memory 204.
[0040] In an example, the data 224 may comprise tire detection training
data 226-1, groove points detection training data 226-2, user data 228, and other data 230. The module(s) 208 may include routines, programs, objects, components, data structures, and the like, which perform particular tasks or implement particular abstract data types. The module(s) 208 may further include modules that supplement applications on the system 102 to measure the tread depth of the tire 400, for example, modules of an operating system. The module(s) 208 may further include modules that implement certain functionalities of the system 102, such as processing the information received by the system 102 from the users, field

representatives, or the channel partners.
[0041] The data 224 serves, amongst other things, as a repository for
storing data that may be fetched, processed, received, or generated by one or more of the module(s) 208. In an example, the user data 228 may be received from the object store 110. The user data 228 may comprise information received from the user corresponding to a request from the user to measure the tread depth of a tire, such as the tire 400, as shown in Figure 4.
[0042] The user seeking to measure the tread depth of the tire 400, may
generate a request for the system 102 to compute a measurement of the tread depth of the tire 400 using any of the user devices 104-1, 104-2, …, 104-n. To receive such a request from the user, in an example embodiment, the system 102 may include a communication module 210. Upon receiving the request from the user to access the system 102, the communication module 210 may prompt the user to upload a target image. The target image may be understood as an input image corresponding to which the user wishes the system 102 to measure the tread depth.
[0043] While receiving the request from the user, the communication
module 210 may require the user to provide other user related data, such as the name of the user, vehicle information, contact number, e-mail address, etc., in addition to the target image. These information received by communication module 210 from the user may be stored as the user data 228 in the central server 108 for further processing by the system 102. The user data 228 may be acquired by the system 102 through the user devices 104-1, 104-2, …, 104-n in response to receiving the request from the user to measure the tread depth of the tire 400.
[0044] In an example, the communication module 210 may provide an
option for the user to upload the target image either by accessing cloud-stored files via public links using any of the user devices 104-1, 104-2, …, 104-n, or directly from locally stored files on any of the user devices 104-1, 104-2, …, 104-n that is being currently used by the user for accessing the

system 102. In another example, the communication module 210 may allow the user to capture the target image via a digital camera of any of the user devices 104-1, 104-2, …, 104-n being used by the user for accessing the system 102.
[0045] For the system 102 to be able to process the request of the user,
it may be necessary for the system 102 to first detect whether the target image provided by the user includes a tire with at least one tread portion. Accordingly, in an example embodiment, the system 102 may include an object detection module 212 that may be adapted to analyze the target image received from the user to detect if the target image is valid, i.e., the target includes a tire 400 with at least one tread portion 402, as shown in Figure 4. The object detection module 212 may discard the target images received from the user that does not include the tire 400 with at least one tread portion 402 and communicate to the user to upload target images correctly.
[0046] In an example embodiment, by using a machine learning model
trained based on the tire detection training data 226-1, the object detection module 212 of the system 102 detects an object of interest, i.e., the tire 400 with at least one tread portion 402, in the target image.
[0047] The tire detection training data 226-1 required for training the
machine learning model to be used by the object detection module 212 to
detect the object of interest in the target image, may be generated by the
training data generation system 232. The system 102 may be
communicatively coupled with the training data generation system 232, as
shown in Figure 2 to access the training data generation system 232.
[0048] In an example, the training data generation system 232 may be a
remote server, similar to the central server 108, that may be accessible to the system 102 through network 106 and used for outsourced computation, data storage, and communication operations, such as generating the tire detection training data 226-1. In some embodiments, the training data generation system 232 may be an application running on the system 102.

In some alternative embodiments, the tire detection training data 226-1 may be generated by the system 102 itself.
[0049] The training data generation system 232 comprises an image
classification module 302 that may be adapted to collect a large dataset of a plurality of random images that may or may not contain the tire 400 with at least one tread portion 402. This dataset is to be used as the tire detection training data 226-1 by the training data generation system 232 to train the machine learning model to be used by the object detection module 212 to detect and classify the object of interest in the target image.
[0050] In an example, the dataset may be created in a database 304
coupled to the training data generation system 232, which may be similar to the object store 110. In some alternative embodiments, the dataset comprising the plurality of random images may be created in the object store 110 and may be accessed by the training data generation system 232. This dataset may include images of different objects including tires, such as the tire 400 as shown in Figure 4, in different poses, lighting conditions, and backgrounds.
[0051] Further, to accurately identify and classify objects within images
contained in the dataset, the dataset may be annotated by labeling each
image with the object class it belongs to. For example, the images
containing the tire 400 may be labeled as "tire image" and the image that
does not contain the tire 400 may be labeled as a "non-tire image". The
image classification module 302 may further extract features from the
images, such as key points or pixels, that may be relevant for the task of
identifying the object of interest, i.e., the tire 400 with at least one tread
portion 402 in the plurality of random images contained in the dataset.
These features may be used as an input to a machine learning model.
[0052] A machine learning algorithm that may be used to train said
machine learning model, including but not limited to a convolutional neural network (CNN), Region-based Convolutional Neural Networks (R-CNNs), Single Shot MultiBox Detector (SSD), may be trained on the dataset to learn

to recognize the tire 400 with at least one tread portion 402 by analyzing the features and labels of the images in the dataset.
[0053] In some embodiments, the training data generation system 232
may employ means to evaluate the performance of the trained machine learning model on a separate dataset. This is done to test the model's ability to generalize to new images and to identify any overfitting issues. Any of the known techniques, such as Split-sample validation, K-fold cross-validation, Bootstrap aggregating (bagging), etc., may be used to evaluate the performance of the machine learning model. Such techniques are well known to those skilled in the relevant art, hence not explained in detail herein.
[0054] Once the machine learning model has been trained and
evaluated, the model may be deployed as a first machine learning engine 216-1 in the object detection module 212 to enable the object detection module 212 to detect the tire 400 with at least one tread portion 402 in the target image provided by the user. This may be done by using the trained machine learning model to classify the target image received from the user, using the features extracted from the target image as input. In some embodiments, the system 102 may be able to process more than one target image at a time if the request received from the user corresponds to providing the measurement of the tread depth in respect of multiple tires at the same time.
[0055] In some embodiments, the machine learning engine 216-1 may be
implemented as a combination of hardware and firmware. In examples described herein, such combinations of hardware and firmware may be implemented in several different ways. For example, the firmware for the module 208 may be processor executable instructions stored on a non-transitory machine-readable storage medium and the hardware for the module may include a processing resource (for example, implemented as either a single processor or a combination of multiple processors), similar to processor 202, to execute such instructions.

[0056] In the present examples, the machine-readable storage medium
may store instructions that, when executed by the processing resource, implement the functionalities of the machine learning engine 216-1. In such examples, the system 102 may include the machine-readable storage medium storing the instructions and the processing resource to execute the instructions. In other examples of the present subject matter, the machine-readable storage medium may be located at a different location but accessible to the system 102 and the processor 202.
[0057] As shown in Figure 4, the tread portion 402 of the tire 400 may
comprise a plurality of grooves 404-1, 404-2, 404-3, 404-4. The tread depth of the tire 400 may be established as a distance between an outermost surface of the tread portion 402 and deepest groove amongst the plurality of grooves 404-1, 404-2, 404-3, 404-4 as measured from a groove base or bottom. Thus, for the system 102 to be able to measure the tread depth of the tire 402, it is necessary that a point be identified on each of the outermost surface of the tread portion 402 and in the deepest groove so that the distance between the outermost surface of the tread portion 402 and the deepest groove may be established.
[0058] Thus, in an example embodiment, once the tire 400 with at least
one tread portion 402 is detected in the target image provided by the user, the target image is fed to a groove points detection module 214. The groove points detection module 214 is configured to detect at least one upper groove point 406-1 and at least one lower groove point 406-2 in the tread portion 402 of the tire 400. The upper groove point 406-1 corresponds to a position on the outermost surface of the tread portion 402, whereas the lower groove point 406-2 corresponds to a location on a ground or bottom of any of plurality of grooves 404-1, 404-2, 404-3, 404-4.
[0059] To detect the upper groove point 406-1 and the lower groove point
406-2 in the tread portion 402 of the tire 400, the groove points detection module 214 comprises a second machine learning engine 216-2 that is also trained following a process as explained above in the context of training of

the first machine learning engine 216-1.
[0060] For example, similar to the first machine learning engine 216-1,
the second machine learning engine 216-2 is also a machine learning model that is trained based on the groove points detection training data 226-2 and deployed in the groove points detection module 214 to detect the upper groove point 406-1 and the lower groove point 406-2 in the tread portion 402 of the tire 400.
[0061] In an example, the image classification module 302 of the training
data generation system may be adapted to collect a large dataset of images that contain test tires having a tread portion with one or more grooves, such as the tire 400 shown in Figure 4. This dataset is to be used as the groove points detection training data 226-2 by the training data generation system 232 to train a machine learning model to be used by the groove points detection module 214 to detect all the upper groove point 406-1 and the lower groove point 406-2 in the tread portion 402 of the tire 400 provided as the target image by the user.
[0062] Similar to the tire detection training data 226-1, the groove points
detection training data 226-2 may be created in the database 304. In some
alternative embodiments, the groove point detection training data 226-2
may be created in the object store 110 and may be accessed by the training
data generation system 232. In some alternative embodiments, the images
that are labeled the as "tire image" by the image classification module 302
may form the groove points detection training data 226-2 to be used as an
input to train the machine learning model to detect the upper and lower
groove points 406-1, 406-2 in the tread portion 402 of the tire 400.
[0063] Further, to ensure that the groove points are correctly and
accurately located in the images of the test tires, the dataset containing the images of the test tires may be annotated by manually identifying a location of the groove points in each image. For instance, the location of key points in the tire 400, such as the upper groove point 406-1 and the lower groove point 406-2, may be annotated, for example, by using a dot in an image

annotation tool, to mark the location of each groove point on the tire 400.
[0064] Next, the images of the test tires and the annotations may be pre-
processed to prepare them for training. This may include resizing the images, normalizing the pixel values, and converting the annotations to a format that may be used as ground-truth data for training the machine learning model.
[0065] Machine learning algorithms or deep learning algorithms that may
be used to train the machine learning model to detect groove points in the tire 400 provided as the target image by the user, including but not limited to a convolutional neural network (CNN), Transformers, a combination of CNNs and Transformers, Region-based Convolutional Neural Networks (R-CNNs), Generative Adversarial Networks (GANs), Recurrent Neural Networks (RNNs), Single Shot MultiBox Detector (SSD). Similar to the machine learning model deployed as the first machine learning engine 216-1, the training data generation system 232 may also evaluate the performance of the machine learning model trained to detect the upper groove point 406-1 and the lower groove point 406-2 in the tread portion 402 of the tire 400.
[0066] Once the machine learning model has been trained and
evaluated, the model may be deployed as the second machine learning engine 216-2 in the groove points detection module 214 to enable the groove points detection module 214 to detect the grooves points, such as the upper groove point 406-1 and the lower groove point 406-2, in a target image which is assessed to include the object of interest, for example, the tire 400 with at least one tread portion 402, as shown in Figure 4, by the object detection module 212.
[0067] After having detected the upper groove point 406-1 and the lower
groove point 406-2 in the tire 400, distance between these two points needs to be computed to measure the tread depth of the tire 400. For this purpose, the system includes a tread depth determination module 218 that is configured to determine a distance between the upper groove point 406-1

and the lower groove point 406-2. The distance calculated between the upper groove point 406-1 and the lower groove point 406-2 may be in the form of pixels and thus may not be useful to the user. Thus, to convert the calculated distance in a known standard scale that may be comprehended by the user, such as a millimeter unit, tread depth determination module 218 may use a fiducial marker, such as an Aruco marker. The procedure of converting the distance obtained between two points via the employment of a machine learning model into a quantifiable unit, such as millimeters, through the utilization of an Aruco marker, constitutes a recognized technique and, as such, need not be described in exhaustive detail within the scope of this patent specification.
[0068] On the deployment of the system 102, in one example, the system
102 may receive a target image. In one example, the system 102 may
receive the target image from the user through the communication module
210. In another example, the system 102 may receive the target image from
another system or device. In one example, the target image may be a
photograph taken by the user. The photograph may be taken based on a
predefined criteria, such as orientation, brightness, width, height, etc.
[0069] On receiving the target image, the object detection module
analyses the target image to detect a tire with at least one tread portion, such as the tire 400, as shown in Figure 4, in the target image.
[0070] On detecting the tire 400 with the tread portion 402 in the target
image provided by the user, the target image is fed to the groove points detection module 214. The groove points detection module 102 of the system 102 then detects the upper groove point 406-1 and the lower groove point 406-2 in the tire 400 detected in the target image provided by the user. Herein, it may be noted that the upper groove point 406-1 and the lower groove point 406-2 is detected in all the plurality of grooves 404-1, 404-2, 404-3, 404-4 of the tire 400.
[0071] Once the upper groove point 406-1 and the lower groove point
406-2 are detected, the tread depth determination module 218 of the system

102 computes the tread depth of the tire 400 by determining the distance
between the upper and lower groove points 406-1, 406-2. In an example,
the distance between the upper and lower groove points 406-1, 406-2, may
be determined by computing a pixel distance between the upper and lower
groove points 406-1, 406-2 by averaging pixel distances between the two
points using Euclidian distance. Thereafter, the tread depth determination
module 218 converts the calculated pixel distance into a known standard
scale that may be comprehended by the user, such as the millimeter unit.
[0072] Although, the present examples have been described considering
only one groove point on each on the outermost surface of the tread portion
402 and the bottom of the one of the one of plurality of grooves 404-1, 404-
2, 404-3, 404-4, it is also possible to use multiple grooves points on each of
the outermost surface of the tread portion and bottom of the grooves and
then take their average to compute the tread depth of the tire 400.
[0073] Once the measurement of the tread depth of the tire is generated
on the real world scale, the measurement is fed to a recommendation module 220. The recommendation module 220, based on the determined tread depth, may generate an action from amongst a plurality of predefined actions to be recommended to the user in response to the user's request for calculating the tread depth of a tire. For example, if the tread depth of the tire is computed to be below a level that is recommended safe for use in the vehicle, the user may be provided a notification, in the form of an e-mail, message, automated voice call, etc., on any of the user devices 104-1, 104-2, …, 104-n through which the request was made by the user for accessing the system 102 to change the tire.
[0074] Therefore, techniques of the present subject matter provide
improved speed and precision in the estimation of the tread depth of a tire without human intervention.
[0075] In an embodiment, in case groove points are not detected in a tire
provided in a target image due to extreme wear of the tread portion, the system 102 may communicate to the user regarding the system's 102

inability to detect the groove in the tire due to extreme wear and hence not able to provide tread depth of the tire.
[0076] In one example, the tread depth of a test tire detected by the
system 102 may be compared to the actual tread depth of the test tire provided as an input to the system 102 in order to determine an accuracy of the training. In a scenario where the detected output is distant from the input provided, a feedback mechanism may be provided to the system 102 in order to increase the accuracy of detection. In a scenario where the detection is inaccurate, the machine learning models employed in the system 102 may be re-trained.
[0077] Figure 5 illustrates a tread depth measurement method 500 (also
referred to as 'method 500'), according to an example of the present subject matter. Although the method 500 may be implemented in a variety of devices, for ease of explanation, the present description of the example method 500 of processing documents is provided in reference to the above-described tread depth measurement system 102 implemented in the network environment 106.
[0078] The order in which the method 500 is described is not intended to
be construed as a limitation, and any number of the described method blocks may be combined in any order to implement the method 500, or an alternative method.
[0079] It may be understood that blocks of the method 500 may be
performed by the depth measurement system 102. The blocks of the
method 500 may be executed based on instructions stored in a non-
transitory computer-readable medium, as will be readily understood. The
non-transitory computer-readable medium may include, for example, digital
memories, magnetic storage media, such as magnetic disks and magnetic
tapes, hard drives, or optically readable digital data storage media.
[0080] Referring to Figure 5, at block 502, a target image is received by
the communication module 210 corresponding to a request from a user for calculating the tread depth of a tire.

[0081] At block 504, the object detection module 212 analyses the target
image received from the user to detect a tire with at least one tread portion, such as the tire 400, as shown in Figure 4. On detecting the tire 400 with the tread portion 402 in the target image provided by the user, the target image is fed to the groove points detection module 214.
[0082] At block 506, the groove points detection module 214 detects the
upper groove point 406-1 and the lower groove point 406-2 in the tire 400 detected in the target image provided by the user. The upper and lower groove points 406-1, 406-2 correspond to a position on an outermost surface of the tread portion 402 and a ground of the tread portion 402 of the tire 400, respectively.
[0083] At block 508, the tread depth determination module 218 computes
the tread depth of the tire 400 by determining the distance between the upper and lower groove points 406-1, 406-2. The distance between the upper and lower groove points 406-1, 406-2, is determined by computing a pixel distance between the upper and lower groove points 406-1, 406-2 by averaging pixel distances between the two points. Thereafter, the tread depth determination module 218 converts the calculated pixel distance into a known standard scale that may be comprehended by the user, such as the millimeter unit.
[0084] It may be understood that the computation of the average pixel
distances is performed exclusively in cases where all upper and lower groove points have been satisfactorily detected. In the event that only a single upper and lower groove point is identified in a tread portion of a tire, the calculation of an average value becomes inadmissible and, as a result, the determination of the distance between the identified upper and lower groove points is the sole basis for the calculation of the tread depth of the tire.
[0085] Thus, the methods and systems of the present subject matter
provide techniques for determining the tread depth of tires that has several benefits. For example, the present technique makes use of machine

learning algorithms to analyze a large amount of images and identify patterns that may not be apparent to the human eye, leading to a more accurate measurement of the tread depth. Also, the machine learning can automate the process of the tread depth measurement, reducing the need for manual inspection. The present technique can also process images of the tires quickly, reducing the time required for inspection. Automating the process with machine learning may reduce labor costs and increase efficiency. Further, the machine learning may be easily scaled to handle large numbers of tires, making it suitable for use in manufacturing plants and other large-scale operations.
[0086] The technique for the measurement of the tread depth discussed
herein is non-destructive, which means it does not damage the tire, which is important for safety and maintaining the integrity of the tire. It also allows preventive maintenance enabled by real-time tire analysis and monitoring, thereby allowing the user to improve the vehicle safety without requiring specialized training. Although implementations of techniques for the identification of concepts in documents have been described in a language specific to structural features and/or methods, it is to be understood that the appended claims are not necessarily limited to the specific features or methods described. Rather, the specific features and methods are disclosed as example implementations for techniques for the identification of concepts in documents.
[0087] Although implementations have been described in a language
specific to structural features and/or methods, it is to be understood that the appended claims are not necessarily limited to the specific features or methods described. Rather, the specific features and methods are disclosed as examples of systems and methods for the tread depth measurement.

I/We Claim:
1. A tread depth measurement system, comprising:
at least one processor;
a communication module coupled to the at least one processor, to receive at least one target image corresponding to a request for calculating a tread depth of a tire;
an object detection module coupled to the at least one processor, to analyze the at least one target image to detect the tire having at least one tread portion, wherein the tread portion comprises one or more grooves;
a groove points detection module coupled to the at least one processor, to detect at least one upper groove point and at least one lower groove point in the target image,
wherein the at least one upper and lower groove points correspond to a position on an outermost surface of the tread portion and a ground of the tread portion, respectively; and
a tread depth determination module coupled to the at least one processor, to determine a distance between the at least one upper and lower groove point, the distance being indicative of the tread depth.
2. The tread depth measurement system as claimed in claim 1, wherein the groove points detection module comprises a second machine learning engine trained based on groove point detection training data to detect at least one upper groove point and at least one lower groove point in the target image, the groove point detection training data comprising manually annotated images of a plurality of test tires, wherein the manual annotation corresponds to a location of one or more upper and lower groove points in at least one of the plurality of test tires.
3. The tread depth measurement system as claimed in claim 1, wherein the tread depth measurement system is communicatively coupled to a training data generation unit, the training data generation unit to:

retrieve images of a plurality of test tires;
present the images for the manual annotation; and
provide the manually annotated images to the second machine learning engine of the tread depth measurement system as the groove point detection training data .
4. The tread depth measurement system as claimed in claim 3, wherein
the training data generation unit further comprises an image classification
module, the image classification module to:
receive a plurality of images; and
classify each of the plurality of images, wherein the classifying comprises identifying images from amongst the plurality of images that include a recognizable tire having at least one tread portion.
5. The system as claimed in any one of claims 3, wherein the training data generation unit is communicatively coupled to a database that stores the manually annotated images of the plurality of test tires.
6. The tread depth measurement system as claimed in claim 1, further comprising a recommendation module coupled to the at least one processor, to, based on the determined tread depth, generate an action from amongst a plurality of predefined actions to be recommended in response to the request for calculating the tread depth.
7. The tread depth measurement system as claimed in claim 1, wherein the tread depth determination module is to compute the tread depth of the tire based on the distance between the at least one upper and lower groove point by determining a pixel distance between the at least one upper and lower groove point by averaging pixel distances between the at least one upper and lower groove point.

8. A tread depth measurement method, comprising:
receiving at least one target image corresponding to a request for calculating tread depth of a tire;
classifying the at least one target image to detect the tire having at least one tread portion, wherein the tread portion comprises one or more grooves;
detecting at least one upper groove point and at least one lower groove point in the target image,
wherein the at least one upper and lower groove points correspond to a position on an outermost surface of the tread portion and a ground of the tread portion, respectively; and
determining a distance between the at least one upper and lower groove point, the distance being indicative of the tread depth.
9. The tread depth measurement method as claimed in claim 8, further comprising interacting with a second machine learning engine trained based on groove point detection training data for detecting at least one upper groove point and at least one lower groove point in the target image, the groove point detection training data comprising manually annotated images of a plurality of test tires, wherein the manual annotation corresponds to a location of one or more upper and lower groove points in at least one of the plurality of test tires.
10. The tread depth measurement method as claimed in claim 8, further comprising generating the training data based on:
retrieving images of a plurality of test tires; presenting the images for the manual annotation; and
providing the manually annotated images to the second machine learning engine as the groove point detection training data.

11. The tread depth measurement method as claimed in claim 8, further comprising, based on the determined tread depth, generating an action from amongst a plurality of predefined actions to be recommended in response to the request for calculating the tread depth.
12. The tread depth measurement method as claimed in claim 8, further comprising computing the tread depth based on the distance between the at least one upper and lower groove point by determining a pixel distance between the at least one upper and lower groove point by averaging pixel distances between the at least one upper and lower groove point.