Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit

DECO 7381    Statement of Work

1 . Purpose

In australia, because of that this area is separated with other continents and its unique weather and ground conditions, there is a lot of unknown and special species, so this is a big work if specialists want to find and research new species as they can,  and the application we design in this project can solve this problem well, it is an mobile application, which is easily downloaded and then used to photograph a new animal then identify if it is already known by now in the current species database or it is failed to be identified after scanning, once the scanning is failed, this picture will be recorded by the system and sent to the taxonomists and zoologists, so that they can targeted identify and name the new species. In addition, users can also share their findings with each other in this application and it is possible that someone knows this species and provides some helpful information to assist the specialists to define this rare or new species. During these functions, users can also share a GPS address under the picture they took, so other people and specialists can know where this species is  active and they can find it and observe it closely.

1. We decide to narrow down the species to animals, which is more specofic to work on and make the application more professional to deal with the problem.

2. Cancel the version of expert, to facilitate the project, we think two versions might be reduntant, and we can actually combine them together in another convinient way.

1.1 Target audience:

Animal enthusiasts and scientists who aim to collect related  data in  existing species.

1.2 Main features:

Identify the animals species

Introduction about animals

Show the location about the animals found by users

Forum page allows users to identify species in pictures uploaded by other users

2. Technologies of each feature

The development model we used is agile development. The overall software architecture is MVC(Model, View, Controller). Therefore, this software will be divided into front end and back end. The front end faces the customer and provides UI interaction. The back end is basically the server responsible for data storage, background logic, and machine learning implementation. The front-end interface can be developed by React Native or Flutter, and the back-end function can be developed by Python. The database can be MySQL.

Functionality will be implemented according to priority.

Core Functions(Implement first):

● Identify animals from photos with CNN image classification.

● Journal log and voting

● Map display animal distribution

Optional functions:

● Track user current location on map using GPS

● Improve performance of image classification:

○ Achieve better generalisation on animal identification (better accuracy on images of different size and scale)

○ Achieve object detection

○ Can identify animals in video

● Community integration & voting

● Ability to identify or suggest a description of the genus of the animals from what was recorded (public interface)

Tools/ technologies/ framework:

1. *Capture photos or videos of species (public interface)

● Machine Learning (CNN(Convolution neural network))

2. Create journal logs for animals  

● Flutter(Flet?)\ React Native

3. Map page to search for special animal

● Flutter(Flet?)\ React Native

● Leaflet JS

4. Use of GPS data to track location of initial finding of species

Allow users to create a journal log

Allows for photos, videos, and notes for the species to be compiled into one post

● GPS

● Open-source API

● Open-source dataset of species

5. Community integration & voting

●  Flutter(Flet?)\ React Native

6. Ability to identify or suggest a description of genus of the animals from what was recorded (public interface)

● Machine Learning:

● CNN(Convolution neural network),

● Database

3. Project Plan  

3.1 Assumptions

(1) We plan to be able to complete the main features we originally planned on time.

(2) In order to realise the main functions of this project, we need to call the collection and recognition of images and videos, as well as the related APIs and technical framework models using GPS positioning. This is the core key in our project. Assume that we have online access to all relevant technical learning resources, open source databases and machine learning models.

(3) All team members are able to attend each team meeting on time and can complete the assigned tasks on time. During the project cycle, when encountering difficult problems, team members can cooperate with each other and communicate and solve problems in a timely manner.

(4) Since our final project requires users to collect images of unknown animals outdoors, we assume that the target users can comply with relevant laws and regulations and ethics.

3.2 Potential risks

4.4 Risk matrix and mitigation

(1)Privacy issues: May relate with user data; Whether photography is allowed.

 Mitigation: Privacy explanation to let users know their data will be used.

(2)Ethical issues: The app may use photographs of animals or humans.

Mitigation: Ask for permission from the committee if necessary.

(3)Data integration: The database can be quite large, and integrating these complex data points can be difficult.

Mitigation: Conduct more research on existing solutions.

4. Team collaboration plan

● Collaboration tools: Using Git for team’s code management

● Communication tools: Teams, Zoom and Wechat

● Meeting time: in class & Friday night

● Meeting content: Show the task process for each member; discuss about existing problem; arrange tasks