Results for "Author: reexre"

2D Game based on my previous post "Boxes". (Yet to be inproved)

Verlet based approach for 2D game physics. 'Original C++ code written by Benedikt Bitterli Copyright (c) 2009 [The code is released under the ZLib/LibPNG license] 'Original C++ code and tutorial available at links http://www.gamedev.net/reference/programming/features/verletPhys/default.asp 'http://www.gamedev.net/reference/articles/article2714.asp 'Forum: 'http://www.gamedev.net/community/forums/topic.asp?topic_id=553845 'Conversion from C++ to Java done by Craig Mitchell Copyright (c) 2010. 'Conversion from "C++ & Java" to VB6 done by Roberto Mior Copyright (c) 2010. 'Joints Handler By Roberto Mior

EVOLUTION BATTLE v5 (Neural Network trained by genetic algorithm) Two teams; a Red one and a Cyan one fight to win battles. Orange one is the "previous battle" best of the Red team. Green one is the "previous battle" best of the Cyan team. (The nearest enemy shot of the "previous battle" best fighter of the Red team is yellow circled) At the End of each battle the team that loses Evolves (The winner team evolves with a Probability of 10%) Each fighter have 11 Inputs and 3 Outputs: Inputs Are: -Enemy Left (0 means Nearest enemy is in front. 0.5 Means Nearest enemy is At 90° Left. 1 Means Nearest Enemy is Back) -Enemy Right (0 means Nearest enemy is in front. 0.5 Means Nearest enemy is At 90° right. 1 Means Nearest Enemy is Back) -Enemy Distance -Enemy Shot Left (As Enemy Left but Refers to "Enemy Shot") -Enemy Shot Right (As Enemy Right but Refers to "Enemy Shot") -Enemy Shot Distance -Availbale Shots -Enemy Relative Orientation -My Velocity -Enemy Velocity -Enemy Relative Moving Direction Outputs Are: -Rocket Left -Rocket Right -Fire Shot (Less than 0.5 do not fire, more than 0.5 Fire) Each fighter have a maximum number of shots that can be fired simultaniously with a given delay. (Now it's set to 1 shot at a time) When fired shot reach the Boundary or hit an enemy it expires. If a fighter hits an enemy with a shot then the fighter Fitness Value is decreased. The hitten enemy is pushed by shot, his fitness increase and it can't perform normal movement for a given time (it's frozen) and his Fitness Value is increased (In this case It's drawn with a cross).(After a fighter has been hitten it becomes "Invisible" to the enemies for a time greater than the time he can't move [frozen]) The team that have the lower avarage fitness wins. (Fitness Value Start from 8000 for each fighters) Fitness is displayed with a line under the fighters. If it is good (low fitness), the line go to Right for the Red team, to Left for the Cyan team. If it is bad (high fitness), the line go to Left for the Red team, to Right for the Cyan team. Evolution is performed by genetic algorithm that modify the genes of fighters. The Genes are the values of a Neural Network wich have the above described Inputs & outputs. Neural Network activity of the "previous battle" best of the Red team (Yellow one) is displaied on a picture. Be patient, the Results of the evolution is noticiable after a long time running (a lot of generations). Program can be stopped and restarted. At the restart (both team) last Evolved fighters are loaded. (Files: Pop1.txt and Pop2.txt) One or both of these files can be deleted to start a new random team population. Enjoy!

Fluid Simulator (Smoothed Particle Hydrodynamics) V5.5 [Update: FASTER] Add Filled Circle - Motion Blur - Blobby2 Draw Style - Customizable Gravity. Old Features: - Add water Filled Rectangle - Blobby Draw Style. Based on this paper: http://www.iro.umontreal.ca/labs/infographie/papers/Clavet-2005-PVFS/pvfs.pdf. Suggestions to make the fluid looks more water-like are appreciated. (line line intersect) (Point distance from line)

Sketch Pictures. Gabor Filter and Strokes. See Screenshot. *** 1.7 -Minor Improvements *** 1.6 -Reported by Robert Bug now Fixed *** 1.5 -Load Settings from previous Rendered Picture (Works only with picture created since this version) *** 1.0 -Faster Gabor Filtering -SOBEL Edge detection -BackGround: Angles By HUE, Angles By Edges Angles, No Strokes: Blurred Source Pic. -No Auto Save -Save As -Hidable/movable panel

Bezier ART. See Screenshot.

This application applies Multipass Bilateral Filter to color Images. Bilateral filtering is an Edge-preserving smoothing filter. This technique extends the concept of Gaussian smoothing by weighting the filter coefficients with their corresponding relative pixel intensities. Pixels that are very different in intensity from the central pixel are weighted less even though they may be in close proximity to the central pixel. This is effectively a convolution with a non-linear Gaussian filter, with weights based on pixel intensities. This is applied as two Gaussian filters at a localized pixel neighborhood , one in the spatial domain, and one in the intensity domain. Some Parameters make result image with a cartoon-like appearance. [ can someone speed up EFF_BilateralFilter Sub? ]

Physical Modelling Synthesis to create sounds of Electric guitar / bass guitar . [ Movable pickups , Frets Slide ] http://www.youtube.com/watch?v=yZhstUIitYc http://www.youtube.com/watch?v=4DbVcyov8nw

NEW:Draw arc of a given Radius tangent to two lines - Line Offset - Fast Atan2 Approx. - Circle Line Intersection - Line Line Intersection - Circle Circle Intersection - Circle from 3 Points - Tangents of 2 Circles - Vector Projection - Vector Reflection - Nearest point on Line - Angle Difference

EVOLVING CREATURES (Roberto Mior) [Upadte 3] - GA better Adaptive Mutation Prob - Added Rotational Velocity for Input (2: >0 and 20 - Particle draw style [Upadte 2] - Smoothed point touching ground function (Sort of pressure) - Smoothed Length Added by Brain - Improved creature physics - X Velocity Air Resistence - Better management of tasks. - Particle System - Increased output links Excursion. - CyclesForScreenFrame 12->16 CyclesForVideoFrame BrainClock [Update 1] - Very Faster - Inputs/outputs Display - SIN Function instead of EXP in NN Activation function - Added Hidden layer with a number of neurons equal to 1.5 times the Number of Outputs - 3 sons per Generation - Better StayOnGround Fitness function - Other ----- This is a typical example of how evolution works. A population of (customizable) Creatures try to Evolve according to some tasks. A Creature is composed by springs and a neural network. There are "intelligent" and "dumb" springs. First ones change their rest length according to the neural network outputs. The other simply react to external forces to mantain their invariable rest lengths. The Creature's Neural Network Inputs are: "No-Brain" Springs Tensions, Points touching ground, Creature Rotation (DX and DY), Height from ground, X and Y Velocities. The Creature's Neural Network Outputs are: "Intelligent" springs rest lengths. EVOLUTION (genetic algorithm): At Each generation there are 3 sons wich parents are "randomly" choosen from the best previous generation creatures. (Some sons mutate their genes with low probability (M)). Then sons are Evaluated. (This 3 sons replace the previous generation worst creatures) TASKS: Creature's evaluation is based on Task(s) to perform. These are: WALK (Distance walked) DON'T JUMP (Minimize avg height from ground) AVG HEIGHT (Average Creature middle point height) MAX HEIGHT (Creature Max Height Reached) VERTICALITY * (Rotate by -/+90°) HORIZONTALITY * (Keep initial orientation)[even 180°] *(Rotation is based on the 2 (left-Right) most external links.) These task can be mixed together. (for some complex tasks/creatures it's a little bit diffcult to find the best proportions) Each Creature performs a score (called "fitness") VIDEOS: http://www.youtube.com/playlist?p=PL52307

This is a typical example of how evolution works. A population of (customizable) Creatures try to Evolve according to some tasks. A Creature is composed by springs, "muscles" and a neural network. There are "intelligent" and "dumb" springs. First ones change their rest length according to the neural network outputs. The other simply react to external forces to mantain their invariable rest lengths. From V6 there are "muscles". Muscle is an angle constrain between two links. The NN outputs determinate even the Angle that they must have. The Creature's Neural Network Inputs are: Vx,Vy & "Pressure" of touching ground Points, Creature Rotation (DX and DY), Height from ground, X Y Velocities and Rotational Velocity. The Creature's Neural Network Outputs are: "Intelligent" springs rest lengths and Muscles rest angles. EVOLUTION (genetic algorithm): At Each generation there are 3 sons wich parents are "randomly" choosen from the best previous generation creatures. (Some sons mutate their genes with low probability (M)). Then sons are Evaluated. (This 3 sons replace the previous generation worst creatures)

My Pencil Sketch Effect 2ND. (see Screenshot) (V4 new background mode) [Gabor Filter - SOBEL (Jähne et al. [1999] variation) ]

RVO. Reciprocal Velocity Obstacle. Crowd Simulation - Video:http://www.youtube.com/watch?v=fN594d80Klg

Draws a sunflower using PHI (Golden Ratio - Golden Number - Fibonacci ). - Video: http://www.youtube.com/watch?v=kWiuFGCI5TU

Google Speak more than 100 character. (and save the MP3) Thanks to Leandro Ascierto