ABSTRACT
The intrusion of technology has penetrated deeper into peopleās lives in twenty first century. The smart phones and other mobile devices have found greater use in many spheres of life. Various applications have been invented to be installed in the smart phones to assist in various works and goals. Despite the increasing use of the smartphones, there is a great scope to use data stored on the devices for wide applications. This leaves room to use smart applications which can provide value to our collected data. As technology is evolving for the ease, simplicity and convenience, our project to assist people to have information about the places they would be visiting. As we are at the peak of the information age, we prefer to get information from the data we feed to our system of the smart-phones. We all want the value of our data, as we are heading towards the age of Artificial Intelligence and Block chain. The main idea of this project is to design a system that will run on most of the smartphones and will be helpful for getting information about the popular tourist destinations by feeding the pictures through the system. We resolve to provide every possible information of a particular place to our user in their palm. The system is processed by the image recognition as a part of Artificial intelligence and information is provided as recognized by the system.
Background theory
The intrusion of technology has penetrated deeper into peopleās lives in twenty first century. The smart phones and other mobile devices have found greater use in many spheres of life. Various applications have been invented to be installed in the smart phones to assist in various works and goals. Despite the increasing use of the smartphones, there is a great scope to use data stored on the devices for wide applications. This leaves room to use smart applications which can provide value to our collected data. As technology is evolving for the ease, simplicity and convenience, we are proposing a project-idea to assist people to have information about the places they would be visiting. In our national context in Nepal, tourism is a very important source of the economy.
Though we are trying to develop many physical infrastructures for tourists inside the country, the aid of technology to provide services to the tourists remains unexplored. One of the major problems tourists face while travelling is lack of information about the places they would seek to visit. Many of them take assistance of local people or professional guide to solve the issue. The main idea of this project is to design a system that will run on most of smart phones and will be helpful for getting information about the visiting places by feeding the pictures through the system. We resolve to give every possible information of that place to our user in their palm. The system will be processed by the image recognition as a part of AI and information will be provided as recognized by the system.
Scope of the project
Currently there is no specific system that is providing information via Image Classification. In this condition, our system has the opportunity to be successful. The scope of this app is primarily focused in Nepal. We decided to operate from selective cities in Kathmandu, Pokhara and Chitwan which are major tourist destination and where information is provided easily. It will interact with different websites like weather, Google maps.
Application
The major application of our app is in the following sectors.Ā
ļ· To provide reliable information about various travel destination throughout the country.
Ā ļ· To act as a PaaS system for the tourism industry.
Ā ļ· To bridge the information gap between travellers and tourism places
The Tourism Industry and Mobile Applications
Having explained the main components in the tourism industry and presenting some relevant data referring to the tourism market in Nepal, the importance of smartphones and mobile applications for this industry will be explained. The expressions cellular (cell) phone, mobile phone, and wireless phone all refer to the same type of voicecentric mobile device that has become the indispensable personal communication tool universally. SmartphoneĀ“s were developed and manufactured by mobile network operators in late 1990s.In recent years, smartphone applications have appeared as a new tool helping travelers create experiences.
Ā Taking into consideration the potential impact of the smartphones and mobile applications, it is important to analyze the usage of smartphone applications in tourism. This dissertation analyses the already existing mobile applications on the tourism market, clarifies which type of consumerās uses these applications during their travel process and predicts some future trends for these new mobile services.Ā
There are several different user groups who use smartphones in a variety of different ways, from business users to media junkies. The literature review shows that the main customers are young men, who use these applications especially to plan their travel during the information phase. But smartphone applications can not only support touristsā information processing activities such as connection and navigation in the tourism consumption stage, but also the activities in the pre-consumption and postconsumption stages.
Data preprocessing and model training:
Images were collected manually through different resolution camera of smart phone and passed through the TF model with MobileNet Architecture and was trained in Google Collaboratory with following configurations.
ā¢ Learning rate: 0.005
ā¢ Input Image size: 224 Ć 224
Images
Classes
Train Images %
Testing Images %Ā
2000
7
70
30
Model freezing
Our aim was to use the trained model in an android application so idea of using the raw model was not viable so we freeze the model i.e. removed the unnecessary nodes form the original Tensor Flow graph, result was an optimized model with reduced size.
Conclusion
The Android app was built for the detection and classification of the landmarks using and getting the information of the landmarks. For the detection we went for offline approach because the frozen mobilenet architecture was quite small. The information was fetched from the wiki server.
References
- “https://www.studytonight.com/mongodb/introduction-to-mongodb,” [Online].
- “Computer Vision: Evolution and Promise,” [Online]. Available: https://cds.cern.ch/record/400313/files/p21.pdf.
- “https://www.projectarm.com/image-recognition-what-it-is/,” 2017. [Online]. Available: https://cds.cern.ch/record/400313/files/p21.pdf.
- A. Rosenfeld, “āPicture Processing by Computerā,” Academic Press, New York, 1969.
- Y. Q. 1. ,. F. R. 1. ,. J. W. 1. a. Q. D. 1. *. Jinmeng Rao 1, “A Mobile Outdoor Augmented Reality Method,” 24 August 2017. [Online].
- X. S. C. W. J. Yu, “Cheong, L.F. A trajectory-based ball detection and tracking algorithm in broadcast tennis video.In IEEE International Conference on Image Processing,” 2004, pp. 1049-1052.
- “https://www.tensorflow.org/,” [Online].Ā
- J. Dacombe, “An introduction to Artificial Neural Networks (with example),” Oct 24, 2017.Ā
- C. S. A. T. D. Erhan, “Deep Neural Networks for Object Detection”.
- S. Saha, “A Comprehensive Guide to Convolutional Neural Networksāāāthe ELI5 way,” Dec 15, 2018.
- “https://www.oreilly.com/library/view/learning-reactnative/9781491929049/ch01.html,” [Online].
- “https://developer.mozilla.org/enUS/docs/Web/JavaScript/About_JavaScript,” [Online].
- “https://www.python.org/doc/essays/blurb/,” [Online].
- “https://www.tutorialspoint.com/nodejs/nodejs_introduction.htm,” [Online].
My relatives always say that I am wasting my time here at web, however I know I
am getting experience all the time by reading such good
articles.