Study with the several resources on Docsity
Earn points by helping other students or get them with a premium plan
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
The course demonstrates concepts related to research methodologies and process; how to conduct a research and present the results of research Topics include Research approaches and methodologies, Writing a research proposal, The literature review, how to gather and analyse data, presentation of findings.
Typology: Study Guides, Projects, Research
1 / 18
This page cannot be seen from the preview
Don't miss anything!
s
Learning Outcomes and Assessment Criteria Pass Merit Distinction LO1 Examine appropriate research methodologies and approaches as part of the research process LO1 & 2 D1 Critically evaluate research methodologies and processes in application to a computing research project to justify chosen research methods and analysis. P1 Produce a research proposal that clearly defines a research question or hypothesis supported by a literature review. P2 Examine appropriate research methods and approaches to primary and secondary research. M1 Evaluate different research approaches and methodology and make justifications for the choice of methods selected based on philosophical/theoretical frameworks. LO2 Conduct and analyse research relevant for a computing research project P3 Conduct primary and secondary research using appropriate methods for a computing research project that consider costs, access and ethical issues. P4 Apply appropriate analytical tools, analyse research findings and data. M2 Discuss merits, limitations and pitfalls of approaches to data collection and analysis. LO3 Communicate the outcomes of a research project to identified stakeholders D2 Communicate critical analysis of the outcomes and make valid, justified recommendations. P5 Communicate research outcomes in an appropriate manner for the intended audience. M3 Coherently and logically communicate outcomes to the intended audience demonstrating how outcomes meet set research objectives.
ASSIGNMENT 1 FRONT SHEET
Student name: Huỳnh Thanh Dương
Centre name: University of Greenwich (Vietnam) - Ho Chi Minh
Tutor: Unit: Computing Research Project
1. Researching about Software 1.1. What tools can help to code into hardware? The languages I code will use Python and C++ and I write those languages through the Arduino IDE and the Python language Arduino is one of the better-known platforms, due in part to the fact that it offers both
hardware and software. The development kit allows users to code in the Arduino language using an integrated development environment (IDE). Currently the cloud system consists of an MQTT broker which enables developers to send messages from one board to another. Projects that have been created using Arduino include a location tracking device for pets, a method to take photos and have them automatically uploaded to popular blogging site Tumblr, and for businesses it’s possible to create a print issue receipt from GitHub onto paper. The Arduino cloud platform will be adding new features over the coming months and are looking for feedback on the current setup. Now seems as good a time as ever to jump on board. 1.2. What is the protocol used to connect to the hardware device? The system is implemented based on Raspberry pi development board in Window environment, supporting SMTP (Simple Mail Transfer Protocol), TCP/IP, HTTP. The Flash File System web server supports dynamically generated files that may include output data from hardware resources. This type of file is called an embedded server page (ESP). 1.3. What tool do you use to create an app on your phone? There are many platforms that you could use to build your IoT application. You can choose any of them, but pay attention to what technologies they support and how much they’ll cost. These are proven platforms for IoT development: HomeKit, Android Things, ThingWorx, Xively, Ubidots, Azure IoT Suite, IBM Watson, Oracle IoT,Kaa Many of these platforms are trying to solve the inconsistency issue, allowing IoT developers to connect devices from different vendors that speak different protocols into one unified system. For example, Google’s IoT platform, called Android Things, now supports devices from a number of manufacturers and can connect to low-energy sensors. Apple’s HomeKit offers a set of approved devices that developers can connect to their IoT systems as well as guidelines for creating IoT software. Choose the hardware and Think of scalability in advance
2. Researching about Hardware 2.1. Which sensors need for project? Raspberry Pi board is a miniature marvel, packing considerable computing power into a footprint no larger than a credit card. The processor at the heart of the Raspberry Pi system is a Broadcom BCM2835 system-on-chip (SoC) multimedia processor. Motion sensor HC-SR501 is a sensor capable of detecting an object moving into the area where the sensor works. The sensor module can adjust the sensitivity thanks to two rheostats, the Sx rheostat, which adjusts the sensitivity of the sensor to suit the requirements of the user. 2.2. Which Wi-Fi microchip need to use for this project? RF Kit BLE ESP32 Wifi Transceiver ESP32-CAM Ai-Thinker Camera manufactured by Ai- Thinker has a compact size with the main processor being the ESP32 module + OV Camera used in image transmission and processing applications. image via Wifi, Bluetooth or IoT applications, the circuit has good processing quality, high durability. 2. 3. Another hardware needs for project.
3.3 Evaluation of a face recognition algorithm:
- % of correct (true) recognition ESP-CAM can determine if two captured faces belong to the same person, with an accuracy rate of 87.25%. When taking the same test, humans answered correctly 87.53% of the time ESP-CAM just got better with FaceNet. On the widely used dataset of Labeled Faces in the Wild (LFW), FaceNet achieved a new record accuracy of 89.63% (0.8963 ± 0.0009). - % of incorrect (false) recognition ESP-CAM can determine if two captured faces belong to the same person, with an accuracy rate of 12.25%. Because the device is still not usable, it is too complicated and the algorithm is still far behind Google or Facebook today with 99%. - How to improve the % of correct recognition? Improved face recognition algorithm: In our work, we acknowledge the fact that the ability of a face recognition algorithm to recognize faces is highly dependent on the precise performance of the feature extraction and comparison stage, which also depends on much on the quality of both the input face image and the training/reference image participating in the face comparison process. To improve the face recognition accuracy of the LBP algorithm, we exploit the following image quality properties of our input and reference face images: brightness, sharpness, noise Improved face recognition algorithm In our work, we acknowledge the fact that the ability of a face recognition algorithm to recognize faces is highly dependent on the precise performance of the feature extraction and comparison stage, which also depends on much on the quality of both the input face image and the training/reference image participating in the face comparison process. To improve the face recognition accuracy of the LBP algorithm, we exploit the following image quality properties of our input and reference face images: brightness, sharpness, noise , resolution, scale and pose, to get the best quality image. show finer details of image features for more precise feature extraction and comparison. , resolution, scale and pose,
to get the best quality image. show finer details of image features for more precise feature extraction and comparison. Plus Image Blending technique will be applied to the training/template face images, then the improved input images will be compared with the improved training images using the LBP algorithm, to yield an improved LBP codes to recognize faces, thus the facial recognition accuracy will be improved compared to the traditional LBP codes without our method. In the second section, the metadata of the recognized facial images such as date and time are automatically extracted to automatically mark attendance of each individual. 3.4 Google Cloud Vision : Face recognition platforms tuned into cloud engines have a massive advantage. If your application runs on Google cloud engine it will be very easy to integrate Google Cloud Vision into your product or application. It boasts many pre-trained models and API outlets to become a power tool for many computer vision programmers. There is an added benefit of availability of AutoML Vision that can help programmes to train customer vision models. Google Cloud Vision has the power to detect multiple faces within an image along with major attributes like emotional state and usage of headband by the person. Google is also reportedly working on adding a facial recognition feature to the suite of tools. 3.5 An Update About Face Recognition on Facebook Facebook has always given you control over whether we use face recognition technology to recognize you in photos. This setting was previously called tag suggestions because Facebook only used face recognition technology to automatically suggest tags of your friends in photos. In December 2017, we introduced a new setting for some people called face recognition. Our face recognition setting lets you manage not only whether Facebook uses face recognition technology on photos of you in order to suggest tags; it provides an easy on or off switch for a broader set of uses of face recognition, such as helping you protect your identity on Facebook. Starting today, people who newly join Facebook or who previously had the tag suggestions setting will have the face recognition setting and will receive information about how it works. The tag suggestions setting, which only controls whether we can suggest that your friends tag you in photos or videos using face recognition will no longer be available. We’ve continued to engage with privacy experts, academics, regulators and people on Facebook about how we use face recognition and the options you have to control it. We’ve made the steps to update your settings clearer and you can opt to leave your setting off right in the notice, as opposed to having to go to a separate screen. If you do nothing, face recognition will remain off for you. There’s also an opportunity to review information about our technology and the features it powers so that you can make the choice that’s right for you. Finally, Facebook’s face recognition technology still does not recognize you to strangers.
using security systems for their stores, but it just stops at a simple level like using cameras to observe. close. Partly because the security system in our country is not well developed, partly because the price is too high. From these necessary needs, the group has come to the decision to choose the topic "Intelligent anti-intrusion alarm system" to research as well as to synthesize and consolidate learned knowledge as well as apply automatic technology. into real life. With this topic, the project team will design a warning model, using ultrasonic, high-frequency and infrared waves. The sensors will send signals to the microcontroller to make phone calls to the user via SIM, email and take a photo alarm when detecting an intrusion.
About Adruino IDE and languages in project: I started learning about Adruino IDE, Python to prepare for project work because I have knowledge of C++ as well as Java. Set a goal every day to complete basic to advanced exercises through youtube. and use my knowledge to practice on tinker cards
Expensive equipment and many complicated devices
Targeting only comes with the door that turns into a smart door unlock system
Activities carried out in a research and development project, idea analysis, writing, data collection, numerical analysis, tutoring meetings, final results, assessments, report writing) and possible time periods: First milestone: Initiating June 23, 2021 - June 30, 20 21
Figure 1 Milestones
[1]. Managementstudyguide.com. 2021. Role of Communication in Presentation and Speech. [online] Available at: [Accessed 5 June 2021]. [2]. Managementstudyguide.com. 2021. Role of Communication in Presentation and Speech. [online] Available at: [Accessed 5 June 2021]. [3]. University of Bradford. 2021. Communication and presentation skills - Developing your skills - University of Bradford. [online] Available at: [Accessed 5 June 2021]. [4]. Author Services. 2021. 5 tips for communicating your research - Author Services. [online] Available at: [Accessed 5 June 2021]. [5]. opscience.iop.org. 2021. ShieldSquare Captcha. [online] Available at: [Accessed 5 June 2021]. [6]. Iopscience.iop.org. 2021. ShieldSquare Captcha. [online] Available at: [Accessed 5 June 2021]. [7]. process, S., 2021. Steps of the research process. [online] Human Kinetics. Available at: [Accessed 20 June 2021]. [8]. S, S., 2021. Difference Between Primary and Secondary Research (with Comparison Chart) - Key Differences. [online] Key Differences. Available at: [Accessed 20 June 2021]. [9]. 2021. [online] Available at: [Accessed 20 June 2021]. [10]. SurveyVitals | Medical Satisfaction Surveys for Quality Improvement. 2021. Cronbach's Alpha Results | SurveyVitals. [online] Available at: [Accessed 20 June 2021].
