The program has a very poor U.I at the moment.
The Apple-identification program has a 5 fail rate at the moment.
Inputting any number into the top box that isn't a single number that has a corresponding image in the 'Maybe' Folder will crash the program
Any input that is not a 16 by 16 pixel PNG will crash the program.
GUI that can take modular input to test.
GUI that displays tested-image
GUI that displays NeuralNet's decision of input image.
(GUI that dspalys 'perfect-apple' based on neural net weights. )
A Weight-based neural net program that can be trained on pixel apple images and identify them from random images.
A Modular design to allow other non-iamge inputs
Adding GUI to Neural Net
*14/02/2021 - 0.2a- GUI update *
Animated Weapon Alpha
To Sum Up: Remove Memory Leak bug for good practice
Add new Enemy Unit type- SoldierGun for fun
Add new Animated Projectile type - target (For when SoldierGun targets player) for variety in combat
Add New Skin- Rambo for a request I got on TikTok
Add New Audio Track - Combat.Wav for player enjoyment
Add Animated player HUD to make game appear more professional
Add Synced music playing - both background music and sound in-game for better immersion
Add debug hitboxes for debug purposes
Add UnitType to keep consistent unit generation
Add DebugTimer for better analytics
Add Corpse Object to allow player to move through bodies
Add Loading screen for aesthetic
13/02/2021{
NeuralNet.cs -> Modify Name from 'Program.cs' This is so new program cs file can run the form object
NeuralNet.cs -> Add static bool 'LastWasApple' to be refrenced if the N.N's decision of an image is needed.
NeuralNet.cs->Main()-> Modify Method to be callable method 'Run Neural Net'
NeuralNet.cs->RunNeuralNet()-> Modify Comments to match
NeuralNet.cs->RunNeuralNet() // NeuralNetForm.cs-> Modify Testing section of the method to be run on button call in NeuralNetForm
NeuralNet.cs->RunNeuralNet()-> Move 'testLayer' to be a static public var so 'TestImage()' can be run using it.
NeuralNetForm.cs -> Add a Form object. This is to have a presentation GUI of the NeuralNet at work.
The intention is it will be able to interate uses of the old console-based 'NeuralNet' program
by making the 'NeuralNet' program a static class that can be called from a form, using a buton and
textbox object to provide values and call the test section through a button press
NeuralNetForm.cs -> OnPaint() -> Add overriding onPaint Method to draw last image searched and a cross or tick
depending on whether the neural net thought it was an apple. This is for visual feedback of the system working.
}
14/02/2021{
NeuralNet.cs->RunNeuralNet()-> Add right Count var for each correct guess program makes with test data.
If more than 1 test data is guessed wrong, reruns program.
}
*14/02/2021 - 0.2a- GUI update *
End Of March -
Cleaner GUI
Simple edge-case handling
More modular user-input / custom input ability
I planned on this project improving my skills in the following:
Understanding A.I and Machine Learning in practice.
Developing custom input/file reading ability
C# and Visual Studio
Making programs more modular and better documented.
Finding apples.
Unzip the 'Images' folder and run the .exe inside.
Type any number between 1-11 and press the button to test whether the Neural Net thinks the given image is an apple. To use custom images replace any image file in the 'Maybe' folder with a 16x16 PNG useing only pure white,red or green pixels.