CITATION

(G. K. Upadhyay et al., 2020)

SUMMARY

This research study aims to develop a futuristic marker-less augmented reality shopping app that providing interactive consumer experience through providing real-time object detection. Another primary goal of this study is to overcome labor costs and consumer’s waiting time in supermarket shopping with the help of the proposed AR app. Further, this research study, analyzing the image tracking of the application by changing different distances to detect the 3D model with the loading time.

This article is useful for AR developers who wish to develop AR apps for identifying the object in a retail market.  This study states that the Vuforia-based augmented reality approach has a 96% accuracy rate on detecting an item or the retail product. This author of this research trying to narrow the space between augmented reality and machine learning. Moreover, his focus is to give more sophisticated service to the customer by joining these emergent technologies. Moreover, In the literature review section, the author describes several futuristic object detection API’s like TensorFlow, Mobile Nets, SSD, and Vuforia.

In the initial stage, the researcher tries on object detection by using tensor flow API. Furthermore, the author states that tensor flow object detection has resulted in less accuracy and fewer fps due to high GPU intensiveness and less efficiency. To overcome these barriers, then the researcher moved to detect the object by using both Single Shot MultiBox Detector (SSD) structure and Mobile Net classifier to overcome these difficulties. The author declared that SSD and Mobile Net classifiers gave better detection accuracy than tensor flow detection. However, due to the limitation of multiple windows for different features by the SSD and Mobile net approach, the author’s focus was on the augmented reality approach.

Initially, the author tries out the marker-based AR. However, due to extra cost and time to create markers, he moves To marker-less augmented reality to detect the retail product in the camera view. Moreover, the application features increased by appending of advertisement video of the product and projecting nutrition facts of the detected product image. Then the actual product images uploaded to the database of Vuforia SDK as target images for the application to assess the rating scores. Animation features for the application arrange by the animation clip, and the script call animator controller controls the animation of the shopping-mart guide. Then includes the video of the advertisement and the product item information.

The author states that python-based object identification, do not provide functionalities that perform by the Vuforia augmented reality technology. Moreover, the author argues that user interface design with AR technology brings the users to feel more real experience and improve today’s AI applications over the other object detection programming. However, the author defines that the Vuforia target images do not always qualify for image tracking. All the uploaded images to the Vuforia SDK’s assessed by the Vuforia target manager according to the rating score range of 0 to 5. Moreover, this study states that markers (images) that do not have enough distribution features and less density will not qualify for image processing.

More importantly, the author state that increasing feature density and appropriate textures with shades can improve the rating stars of the image target manager provided by the Vuforia engine. Also, the author declaring that marker-based AR implementation for the retail industry will bring additional cost to the production of markers. Furthermore, the researcher stating that the marker-less technique is a better approach that will bring a solution to cut down the extra cost of production of markers for a different product.

After implementing the research project, the author concluded that all the test results had the expected result, and all the research’s function and performance gave what the user expected.

CITATION

(Song et al., 2019)

SUMMARY

The objective of this research paper is to explain how augmented reality try-on experience supports the future shopping decision on the consumers on buying the product considering marker base swiss-watch app.

The author of this study analyzes the psychological states and the consumers’ behaviors trying on AR applications. The result of the author’s analysis was used as the methodology to conclude the research purpose. This research paper uses the AR try-on application implemented by the Swiss watch company to measure the consumers’ psychological behaviors.

Furthermore, the author of this application-defined AR try-on watch application was brought more profit and enhancing on more consumers to the swiss watch company. The author of this study uses the theoretical approach of EE and SPC algorithm to determine how the AR experience can make the effect of the consumer decision. According to the analysis of the above algorithm, the author state that from AR try-on application who try on the virtual shopping item can feel ownership of that item before buying. That brings the customer to reach the item more closely to support their decision making. More, further, this research work helps my current research study determine the benefit that I can obtain by the marker-base AR solution in the retail industry.

CRITIQUE

According to the research article, the author states that scientific research was analyzed by the mobile application, which was available at that time. Further, this study considers only the essential psychological and behaviors of consumers. Also, this study considers only the online customer and asking for decision comfort about purchasing only the luxury type of watch. Further, the current research model considers the only total of 99 responses for the experiment, And we feel that the researcher should extend the sample size more to get the correct result and should consider the respondent’s age, occupation, and cultural attributes.

CITATION

(Syihabudin et al., 2020)

SUMMARY

The primary objective of this study is to identify the benefit of marker-based augmented reality (AR) application by developing a smartphone AR application to visualize 3D animals in real-world environments. Furthermore, this study analyses the efficiency of the angle detection algorithm that supports detecting the marker-based object. Furthermore, this research paper, concluding that the marker-based tracking approach was showing a faster detection, and the author of this study analyzing the implemented the 3D animal application with different of 300 angles and declaring that the application has detected 54.17% registered image target.

On the other hand, this research paper is one of the useful articles for researchers who want to try out developing AR applications because this study brings up knowledge of developing AR applications from scratch, such as data gathering (research method), designing, implementation, and testing. In the stage of data gathering, this article provides how the development team gather the 3D models of the animals by adding information, sound, and rotation. In the design phase, the researcher describes how the application was design the screen for start menus, help screens, about screen, and 3D views design.

In the implementation phase researcher describe how the AR marker should be required to design to capture the 3D model on top of it. Moreover, this article gives facts of development information of the software tools and SDKs that use to implement the AR experience on top of the design marker. In the testing stage, the author provides a bunch of testing results and the application screens that conclude the evidence on how the application looks like in different mobile devices.

This research paper indicates the software tools, SDKs, and libraries used to develop an AR application. The author highlighted the “Unity Asset Store,” which contains a library of free and commercial 3D object models built into the Unity editor. In this study, the researcher indicated predefined virtual animals’ objects imported from the Unity asset store. The screen of the start menu, help menu, and other navigation screens developed using 2D shapes in the Unity 3D game object tools bar.

In this research study, the application marker designed by black and white illustrations with a thick black border and white background. The AR  marker was validated by uploading to the Vuforia engine database and given an all-stars rating of 1 to 5. Once the application developed, the app tested using different mobile devices against targeting the marker by staying in different distances and other lighting conditions. This test result indicates how well the AR marker is getting triggered in a different environment.

CRITIQUE

The author of this article states that the marker-based AR is a faster image detection technique. From his research, the author declaring that 54% of registered marker detected without any issue.  However, According to the article (A. Alvaro-Tordesillas et al., 2019), it brings the development process much more comfortable without a marker. Also, the author (Blanco-Pons et al., 2019) states that marker-based tracking makes an easy way to identify the target image, but different design markers can bring out extra work compared to marker-less tracking.

Traditionally, to visualize the AR experience, the consumers should wear mobile devices or head-mounted wearable devices such as Microsoft Hololenz. Due to that, consumers always need to wear some instruments to get into the AR reality. Moreover, this may disrupt the customer’s attention from the physical environment. However, in this research paper, the author aims to avoid these barriers by producing AR experience to the customers without disturbing the natural customer experience by implementing the concept of augmented spatial reality (SAR). So this article research is beneficial for the AR developers who try to develop augmented spatial reality(SAR) experience to the customers’ realistic view.

This researcher of this study user SAR experience conceptually divided into two user scenarios, self-interaction and book interaction. In the Self-interaction, the system can recognize user interaction when the user approaches the book store and guides them by providing digital information for browsing books, like view pricing and book recommendation on the user’s selected book.  In the book interaction, the specific book interacts with the user according to the book’s instruction. (book talking).

To track the user’s position, the author state that this research used cross-platform skeletal tracking and gesture recognition software called Nuitrac. Moreover, the researcher declaring that to map the physical object in real-world content, they used the Vuforia marker-based image recognition technique. Moreover, They placed an RGB camera direct to the view of the pages of the book.

CITATION

(A. Alvaro-Tordesillas et al., 2019)

SUMMARY

This research paper aims to develop a mobile application for a museum environment to provide better user interaction by giving more experience by projecting a 3D sculptural object from two-dimensional images using augmented reality using markerless experience. Further, this article’s author defined the object’s recognition track with the help of Vuforia SDK, an augmented reality development kit (SDK). Usually, AR applications implemented using marker base tracking due to a tight angle detection algorithm. (Syihabudin et al., 2020). However, The novelty and particularity of this research are to propose a method that eliminates the marker base image tracking by converting the real object as a marker. This research proposed a method to eliminate the marker that acts as intermediation between the surface and 3D objects.

The author of this article states that after developing the application, the scientific interest test locates in the Jorge Oteiza Museum Foundation in Alzuza (Pampoloan) in Spain.  The scientific test approaches by setting the object’s distance, setting different camera height to the item, and projecting the virtual object in other lighting conditions.

According to the author of this report, the scientific test approaches in the main two ways,

1. Simple anamorphosis – changing the object’s distance, setting different camera height to the item, and projecting the virtual object in other lighting conditions. The test has experimented between 50 and 100 cm. Moreover, the hight is taken to the account from 120 cm from the ground level with 3.43/5.14 mm camera lenses. Moreover, the author states that from this method, all the result was given a positive response by detecting image and producing the virtual content.

1. Analyzing marker-based image identification by building a simple 3D object, such as boxes in real-time, using an augmented reality software development kit (SDK), and testing them with different items, according to four characteristics, color, size, geometrically and with including additional objects on to it. From the result, the researcher state that the marker image recognizes chromatic differences (hue, saturation, and brightness), and all the impact were positive. However, when the researcher tests the marker by introducing another object next to it, the result becomes negative.

From the result of the scientific interest, the author of this report concludes that AR experience with marker base can simplify by eliminating the intermediate marker by converting the real object as a marker.

CRITIQUE

In this study, after analyzing the result of the experiment of marker-less image tracking, the author concludes that whenever one object target introduces closer to another object target, then the image identification is given negative results. Also, he does not provide any more comprehensive information about the research in different parameters.

CITATION

(Purnomo et al., 2018)

SUMMARY

This research paper aims to develop a mobile application for an animal’s museum environment to project ancient objects using visitors camera angels with the help of Vuforia Augmented reality markerless object recognition. The implementation content includes virtual 3D animal objects that project in real-world exhibition ancient objects and related animal’s information with their audio sound clips.  Furthermore, this study brings scientific interest to the AR technology by providing results on the success rate of the AR detection object in a different environment.

According to the research study of (Blanco-Pons et al., 2019), in marker-based AR, the target image should be registered in the system (Vuforia), and it should have enough key points to save into the database of the system (Vuforia) to identify the image target in real-world content.  The researcher of this study state that one of the significant issues in the AR system is the challengable accuracy rate in registering 3D objects in real-world content. Due to that reason, the author, in this paper, tries to research on accuracy level on the ability of AR detecting on the marker-less exhibition object tracking approach.

This research was undertaken to detect the ability of markers detection by changing of including distance, angle, shadows, lighting, and movement/motion. In this research, the 1^st experiment was done by the change of angles according to the user camera view targeting  73 3D objects (between 0 -360 angle and 0 – 180 angle). Moreover, each object’s lightning is given a range of between 8 Lux to 100 Lux.    The author of this study state that the beginning of the experiment with 54% keypoint did not recognize the AR app. However, after optimizing the features of the exhibition object, Vuforia image registration given the rating above the value of 3 ranges from 1 to 5 range.

From the research result, the author concludes that angle position horizontally range of 0 to 25, 3D markerless detected, and give a positive result. However, in this research, the author stating that lighting can give vulnerable output in 3D AR marker-ess approach. In this study, the researcher state that changing lighting condition is much more crucial in this project because museum showroom lighting and outside the room lighting can impact on the object AR markerless object tracking on exhibition items.

CRITIQUE

One of the popularity weaknesses of marker-less AR is the difficulty of image recognization. Same as in this research, the author defining that from the research result, the lighting condition with the value of 10Lux, 30Lux, and 100 Lux given better results on tracking the 3D object in this research. However, the author declaring that reduction of the intensity of the light too large, doest give the effect in image recognization as expected.

Further, the author of this report declares that AR 3D markerless can detect the 3D object, only having a deviation angle of 25⁰, which is a tilt angle to 25^0.   Also, he concludes that 3D AR markerless is not recognizable while outside from the above angle range.

CITATION

(Pantano et al., 2017)

SUMMARY

This research aims to investigate the cross-country impact of augmented reality technology on consumer’s actions within e-commerce by comparing two marker-less virtual try-on (a smart mirror for virtual glasses) mobile apps implemented by Italian and german vendors. The author of this research analyzed the behaviors and motives responses of 318 young Italian and German consumers.

In this article, the researcher analyzed the consumers’ willingness to buy the markerless try-on app, based on the technology acceptance model (TAM) and according to the technical aspects (kind of the information, appreciative quality, interactivity).  Moreover, the mean values of the two samples declare as following, German samples as 7.81 (already users and non-users) and the Italian sample as (10.14). From the result was taken by two cultural samples, the author state that both consumers in two groups are given a positive attitude and consider the AR application (smart try-on marker-less application) as a powerful technology to support their decision-making process and behaviors. Moreover, essentially, the author declared that participants thanks to technical characteristics.

This research paper is chosen as a literature review article for the current undergoing going research study because to analyzes the result of this paper providing evidence on the effect on marker-less image processing by the retail industry across the different cultures.

CRITIQUE

This research article investigated the experiment by focusing only on a specific age range (between 20 and 30). According to the U.S inter usage presentation, most of the people in the world age 30-49 years old using 97%  internet features. However, the author of this research focuses only on age around 20 to 30 only.

Moreover, according to the article of (Purnomo et al., 2018) state that, marker-less AR technology having weakness result in less lighting condition. In this research paper, the researcher does not analyze customer behaviors by setting up in different lighting environments.

Due to these points, we can conclude that researcher is a little bias on proving his idea in this research.

CITATION

(Kim & Kim, 2014)

SUMMARY

The objective of this research study is to implement a markerless augmented reality application that runs in smartphone devices, which help to retrieve the information of advertisements (food restaurants).  The author of this article state that this study brings connectivity to the backend database.

This article declares that the application implemented using Vuforia SDK v2.6  and  Unity, and the database management system used is MySQL. and data transmission between the app and the backend web data service by using PHP and XML language

In this study, the restaurant leaflet’s image and outdoor signboard registered as the image target in the Vuforia database. Moreover, the 3D model with the 2D image canvas for (selling food and booking seat details) developed in Unity 3D when an advertising image matches the 3D model in unity. The author states that the 3D object with the 2D user interface was augmented in the advertising image.

CRITIQUE

The researcher in this article does not illustrate how the application is tested with the expected results. Furthermore, the author does not provide any scientific interest in this research article. The figure also images not enough to determine the application process. According to the article (Sadeghi-Niaraki & Choi, 2020), determine that the AR markerless tracking in image targets can deviate under indoor and outdoor conditions. According to that statement, we can conclude that this article’s researcher considers the implementation of the application but does not include any information about the accuracy rate of the markerless tracking.

CITATION

(Syahputra et al., 2020)

SUMMARY

This article, published as an academic journal, calls “Augmented Reality Virtual House Model Using ARCore Technology Based on Android.”  The author of this article tries to research the accuracy of surface tracking by creating an AR application using ARCore markerless tracking SDK, which is developed by Google.

This research paper includes developing a AR 3D virtual object model and displaying it in a real-world environment by detecting flat surfaces. This article explains the two types of behaviors of  AR applications of marker tracking and markerless tracking. According to this article,  marker tracking requires a specific marker to detect on camera and then to display a AR 3D model top of it. However, markerless tracking does not need a marker to remain on the frame of the camera is no longer needed.

In this research paper, the author argues that the marker on augmented reality has limitations and less interactive. The author points out that the marker should need to capture within the range of the camera, so the user is not able to move freely. However, markerless AR having more interactive can display even multiple objects at once and can display the virtual object on top of any surface rather than requiring a marker to capture on camera.

In this paper, the author research the accuracy and stability of AR Core detection on a flat surface to display a virtual object top of the surface. The researcher implemented an AR application using AR Core and test the accuracy and stability of popping up the 3D model of the top of the real-world surface like surface corner room, corridor, wall room, and stairs border. This research had done by the testing number of 3D points that display when the user points out the phone camera to different surfaces.

CRITIQUE

Based on the test result from the research, the author concludes that AR Core plane detection can detect the flat surfaces that having different textures. However, textures that are having single colores having difficulty in determining the surface, for example, while color surface and darkroom having difficulty producing 3D points on the surface.  Further, according to the article of (Blanco-Pons et al., 2019) marker-less image tracking approach giving un-stable results due to the angle of detection. But in this research, the author does not provide enough evidence on testing results.

CITATION

(Sadeghi-Niaraki & Choi, 2020)

SUMMARY

This research paper provides a complete overview of AR marker-less registration of the image tracking approach. Further, the author of this article research on AR marker-less application that is focusing on Geospatial information system (GIS) that works as a conceptualized framework that has the capability and examines the spatial and geographic data. Moreover, the primary goal of this research is to review the literature on the markerless image tracking technique used by the existing indoor and outdoor applications and finally try to declare which approach is more suitable for this purpose.

The seamless and accurate result taken from outdoor (environmental applications) and indoor (health applications) applications, the author analyzed them based on two techniques, such as sensor-based and positioning technique, vision-based methods.

The sensor-based approach is experiment with image tracking by changing the orientation and movement by using an accelerometer and a gyroscope based on the location and the camera’s orientation. According to the result was taken from the sensor-based approach, the researcher declaring that sensor-based tracking provided several sources of errors resulting in a low level of tracking accuracy.

However, The author state that the vision-based approach, which experiments process by the edge-based method, no-model technique, and template matching method, was given a more accurate and reliable tracking result comparison to a sensor-based approach.

CRITIQUE

In this study, the author concludes that point base detection on edges under the vision method given more reliable results every time. However, according to the article (Blanco-Pons et al., 2019), the author concludes, point-based marker-less image tracking does not always give on positive results. Moreover, the surface or marker does not have enough features to identify the target always low on detecting the target object.

CITATION

(Kharroubi et al., 2020)

SUMMARY

The article of this research paper proposing a cloud web-based technology that stores the 3D points to detect the features of the image targets called web-based-marker-less mobile augmented reality. The reason for choosing this research paper for my research is to know how the future is going to be in AR. This research paper is investigating the creating of a cloud service that can be connected through AR  mobile applications web browsers to test the performance of the AR system by using 29 million 3D points. However, the author state that this hope should need to research more on due to delay related to the quality of network connection. And the high power consumption for the built-in portable sensors. Moreover, this research paper showing different web-AR approaches implemented by the current mobile system. According to the research study, the 3D massive point cloud for the image identification develop by using  Three.js, which is a javascript library  (which can create animated 3D models)  and the features of VR and  AR use by the API called WebXR. According to the author state that this technology can be more benefitted to the AR users who used AR experience as a mobile app because all the computational processing happened in the cloud-based, not in their phone devices.

CRITIQUE

According to this article, the author state that the challenge of the web-based marker-less Augmented reality, network delay, and battery power is taken when the marker tracking is processing. Also, the author should be aware of the security aspect and the internet connection while developing web-based AR.