annotate org/vision.org @ 472:516a029e0be9

complete first draft of hearing.
author Robert McIntyre <rlm@mit.edu>
date Fri, 28 Mar 2014 18:14:04 -0400
parents 3401053124b0
children 819968c8a391
rev   line source
rlm@34 1 #+title: Simulated Sense of Sight
rlm@23 2 #+author: Robert McIntyre
rlm@23 3 #+email: rlm@mit.edu
rlm@38 4 #+description: Simulated sight for AI research using JMonkeyEngine3 and clojure
rlm@34 5 #+keywords: computer vision, jMonkeyEngine3, clojure
rlm@23 6 #+SETUPFILE: ../../aurellem/org/setup.org
rlm@23 7 #+INCLUDE: ../../aurellem/org/level-0.org
rlm@23 8 #+babel: :mkdirp yes :noweb yes :exports both
rlm@23 9
ocsenave@264 10 * JMonkeyEngine natively supports multiple views of the same world.
ocsenave@264 11
rlm@212 12 Vision is one of the most important senses for humans, so I need to
rlm@212 13 build a simulated sense of vision for my AI. I will do this with
rlm@306 14 simulated eyes. Each eye can be independently moved and should see its
rlm@212 15 own version of the world depending on where it is.
rlm@212 16
rlm@306 17 Making these simulated eyes a reality is simple because jMonkeyEngine
rlm@306 18 already contains extensive support for multiple views of the same 3D
rlm@218 19 simulated world. The reason jMonkeyEngine has this support is because
rlm@218 20 the support is necessary to create games with split-screen
rlm@218 21 views. Multiple views are also used to create efficient
rlm@212 22 pseudo-reflections by rendering the scene from a certain perspective
rlm@212 23 and then projecting it back onto a surface in the 3D world.
rlm@212 24
rlm@218 25 #+caption: jMonkeyEngine supports multiple views to enable split-screen games, like GoldenEye, which was one of the first games to use split-screen views.
rlm@212 26 [[../images/goldeneye-4-player.png]]
rlm@212 27
ocsenave@264 28 ** =ViewPorts=, =SceneProcessors=, and the =RenderManager=.
rlm@306 29 # =ViewPorts= are cameras; =RenderManger= takes snapshots each frame.
ocsenave@264 30 #* A Brief Description of jMonkeyEngine's Rendering Pipeline
rlm@212 31
rlm@213 32 jMonkeyEngine allows you to create a =ViewPort=, which represents a
rlm@213 33 view of the simulated world. You can create as many of these as you
rlm@213 34 want. Every frame, the =RenderManager= iterates through each
rlm@213 35 =ViewPort=, rendering the scene in the GPU. For each =ViewPort= there
rlm@213 36 is a =FrameBuffer= which represents the rendered image in the GPU.
rlm@151 37
rlm@306 38 #+caption: =ViewPorts= are cameras in the world. During each frame, the =RenderManager= records a snapshot of what each view is currently seeing; these snapshots are =FrameBuffer= objects.
ocsenave@265 39 #+ATTR_HTML: width="400"
ocsenave@272 40 [[../images/diagram_rendermanager2.png]]
ocsenave@262 41
rlm@213 42 Each =ViewPort= can have any number of attached =SceneProcessor=
rlm@213 43 objects, which are called every time a new frame is rendered. A
rlm@306 44 =SceneProcessor= receives its =ViewPort's= =FrameBuffer= and can do
rlm@219 45 whatever it wants to the data. Often this consists of invoking GPU
rlm@219 46 specific operations on the rendered image. The =SceneProcessor= can
rlm@219 47 also copy the GPU image data to RAM and process it with the CPU.
rlm@151 48
ocsenave@264 49 ** From Views to Vision
ocsenave@264 50 # Appropriating Views for Vision.
rlm@151 51
ocsenave@264 52 Each eye in the simulated creature needs its own =ViewPort= so that
rlm@213 53 it can see the world from its own perspective. To this =ViewPort=, I
rlm@306 54 add a =SceneProcessor= that feeds the visual data to any arbitrary
rlm@213 55 continuation function for further processing. That continuation
rlm@213 56 function may perform both CPU and GPU operations on the data. To make
rlm@213 57 this easy for the continuation function, the =SceneProcessor=
rlm@306 58 maintains appropriately sized buffers in RAM to hold the data. It does
rlm@218 59 not do any copying from the GPU to the CPU itself because it is a slow
rlm@218 60 operation.
rlm@214 61
rlm@213 62 #+name: pipeline-1
rlm@213 63 #+begin_src clojure
rlm@113 64 (defn vision-pipeline
rlm@34 65 "Create a SceneProcessor object which wraps a vision processing
rlm@113 66 continuation function. The continuation is a function that takes
rlm@113 67 [#^Renderer r #^FrameBuffer fb #^ByteBuffer b #^BufferedImage bi],
rlm@306 68 each of which has already been appropriately sized."
rlm@23 69 [continuation]
rlm@23 70 (let [byte-buffer (atom nil)
rlm@113 71 renderer (atom nil)
rlm@113 72 image (atom nil)]
rlm@23 73 (proxy [SceneProcessor] []
rlm@23 74 (initialize
rlm@23 75 [renderManager viewPort]
rlm@23 76 (let [cam (.getCamera viewPort)
rlm@23 77 width (.getWidth cam)
rlm@23 78 height (.getHeight cam)]
rlm@23 79 (reset! renderer (.getRenderer renderManager))
rlm@23 80 (reset! byte-buffer
rlm@23 81 (BufferUtils/createByteBuffer
rlm@113 82 (* width height 4)))
rlm@113 83 (reset! image (BufferedImage.
rlm@113 84 width height
rlm@113 85 BufferedImage/TYPE_4BYTE_ABGR))))
rlm@23 86 (isInitialized [] (not (nil? @byte-buffer)))
rlm@23 87 (reshape [_ _ _])
rlm@23 88 (preFrame [_])
rlm@23 89 (postQueue [_])
rlm@23 90 (postFrame
rlm@23 91 [#^FrameBuffer fb]
rlm@23 92 (.clear @byte-buffer)
rlm@113 93 (continuation @renderer fb @byte-buffer @image))
rlm@23 94 (cleanup []))))
rlm@213 95 #+end_src
rlm@213 96
rlm@273 97 The continuation function given to =vision-pipeline= above will be
rlm@213 98 given a =Renderer= and three containers for image data. The
rlm@218 99 =FrameBuffer= references the GPU image data, but the pixel data can
rlm@218 100 not be used directly on the CPU. The =ByteBuffer= and =BufferedImage=
rlm@219 101 are initially "empty" but are sized to hold the data in the
rlm@306 102 =FrameBuffer=. I call transferring the GPU image data to the CPU
rlm@213 103 structures "mixing" the image data. I have provided three functions to
rlm@213 104 do this mixing.
rlm@213 105
rlm@213 106 #+name: pipeline-2
rlm@213 107 #+begin_src clojure
rlm@113 108 (defn frameBuffer->byteBuffer!
rlm@113 109 "Transfer the data in the graphics card (Renderer, FrameBuffer) to
rlm@113 110 the CPU (ByteBuffer)."
rlm@113 111 [#^Renderer r #^FrameBuffer fb #^ByteBuffer bb]
rlm@113 112 (.readFrameBuffer r fb bb) bb)
rlm@113 113
rlm@113 114 (defn byteBuffer->bufferedImage!
rlm@113 115 "Convert the C-style BGRA image data in the ByteBuffer bb to the AWT
rlm@113 116 style ABGR image data and place it in BufferedImage bi."
rlm@113 117 [#^ByteBuffer bb #^BufferedImage bi]
rlm@113 118 (Screenshots/convertScreenShot bb bi) bi)
rlm@113 119
rlm@113 120 (defn BufferedImage!
rlm@113 121 "Continuation which will grab the buffered image from the materials
rlm@113 122 provided by (vision-pipeline)."
rlm@113 123 [#^Renderer r #^FrameBuffer fb #^ByteBuffer bb #^BufferedImage bi]
rlm@113 124 (byteBuffer->bufferedImage!
rlm@113 125 (frameBuffer->byteBuffer! r fb bb) bi))
rlm@213 126 #+end_src
rlm@112 127
rlm@213 128 Note that it is possible to write vision processing algorithms
rlm@213 129 entirely in terms of =BufferedImage= inputs. Just compose that
rlm@273 130 =BufferedImage= algorithm with =BufferedImage!=. However, a vision
rlm@213 131 processing algorithm that is entirely hosted on the GPU does not have
rlm@306 132 to pay for this convenience.
rlm@213 133
ocsenave@265 134 * Optical sensor arrays are described with images and referenced with metadata
rlm@214 135 The vision pipeline described above handles the flow of rendered
rlm@214 136 images. Now, we need simulated eyes to serve as the source of these
rlm@214 137 images.
rlm@214 138
rlm@214 139 An eye is described in blender in the same way as a joint. They are
rlm@214 140 zero dimensional empty objects with no geometry whose local coordinate
rlm@214 141 system determines the orientation of the resulting eye. All eyes are
rlm@306 142 children of a parent node named "eyes" just as all joints have a
rlm@214 143 parent named "joints". An eye binds to the nearest physical object
rlm@273 144 with =bind-sense=.
rlm@214 145
rlm@214 146 #+name: add-eye
rlm@214 147 #+begin_src clojure
rlm@215 148 (in-ns 'cortex.vision)
rlm@215 149
rlm@214 150 (defn add-eye!
rlm@214 151 "Create a Camera centered on the current position of 'eye which
rlm@338 152 follows the closest physical node in 'creature. The camera will
rlm@338 153 point in the X direction and use the Z vector as up as determined
rlm@338 154 by the rotation of these vectors in blender coordinate space. Use
rlm@338 155 XZY rotation for the node in blender."
rlm@214 156 [#^Node creature #^Spatial eye]
rlm@214 157 (let [target (closest-node creature eye)
rlm@338 158 [cam-width cam-height]
rlm@338 159 ;;[640 480] ;; graphics card on laptop doesn't support
rlm@338 160 ;; arbitray dimensions.
rlm@338 161 (eye-dimensions eye)
rlm@215 162 cam (Camera. cam-width cam-height)
rlm@215 163 rot (.getWorldRotation eye)]
rlm@214 164 (.setLocation cam (.getWorldTranslation eye))
rlm@218 165 (.lookAtDirection
rlm@338 166 cam ; this part is not a mistake and
rlm@338 167 (.mult rot Vector3f/UNIT_X) ; is consistent with using Z in
rlm@338 168 (.mult rot Vector3f/UNIT_Y)) ; blender as the UP vector.
rlm@214 169 (.setFrustumPerspective
rlm@338 170 cam (float 45)
rlm@338 171 (float (/ (.getWidth cam) (.getHeight cam)))
rlm@338 172 (float 1)
rlm@338 173 (float 1000))
rlm@215 174 (bind-sense target cam) cam))
rlm@214 175 #+end_src
rlm@214 176
rlm@214 177 Here, the camera is created based on metadata on the eye-node and
rlm@273 178 attached to the nearest physical object with =bind-sense=
rlm@214 179 ** The Retina
rlm@214 180
rlm@214 181 An eye is a surface (the retina) which contains many discrete sensors
rlm@470 182 to detect light. These sensors can have different light-sensing
rlm@470 183 properties. In humans, each discrete sensor is sensitive to red, blue,
rlm@470 184 green, or gray. These different types of sensors can have different
rlm@470 185 spatial distributions along the retina. In humans, there is a fovea in
rlm@470 186 the center of the retina which has a very high density of color
rlm@470 187 sensors, and a blind spot which has no sensors at all. Sensor density
rlm@470 188 decreases in proportion to distance from the fovea.
rlm@214 189
rlm@214 190 I want to be able to model any retinal configuration, so my eye-nodes
rlm@214 191 in blender contain metadata pointing to images that describe the
rlm@306 192 precise position of the individual sensors using white pixels. The
rlm@306 193 meta-data also describes the precise sensitivity to light that the
rlm@214 194 sensors described in the image have. An eye can contain any number of
rlm@214 195 these images. For example, the metadata for an eye might look like
rlm@214 196 this:
rlm@214 197
rlm@214 198 #+begin_src clojure
rlm@214 199 {0xFF0000 "Models/test-creature/retina-small.png"}
rlm@214 200 #+end_src
rlm@214 201
rlm@214 202 #+caption: The retinal profile image "Models/test-creature/retina-small.png". White pixels are photo-sensitive elements. The distribution of white pixels is denser in the middle and falls off at the edges and is inspired by the human retina.
rlm@214 203 [[../assets/Models/test-creature/retina-small.png]]
rlm@214 204
rlm@214 205 Together, the number 0xFF0000 and the image image above describe the
rlm@214 206 placement of red-sensitive sensory elements.
rlm@214 207
rlm@214 208 Meta-data to very crudely approximate a human eye might be something
rlm@214 209 like this:
rlm@214 210
rlm@214 211 #+begin_src clojure
rlm@214 212 (let [retinal-profile "Models/test-creature/retina-small.png"]
rlm@214 213 {0xFF0000 retinal-profile
rlm@214 214 0x00FF00 retinal-profile
rlm@214 215 0x0000FF retinal-profile
rlm@214 216 0xFFFFFF retinal-profile})
rlm@214 217 #+end_src
rlm@214 218
rlm@214 219 The numbers that serve as keys in the map determine a sensor's
rlm@214 220 relative sensitivity to the channels red, green, and blue. These
rlm@218 221 sensitivity values are packed into an integer in the order =|_|R|G|B|=
rlm@218 222 in 8-bit fields. The RGB values of a pixel in the image are added
rlm@306 223 together with these sensitivities as linear weights. Therefore,
rlm@214 224 0xFF0000 means sensitive to red only while 0xFFFFFF means sensitive to
rlm@214 225 all colors equally (gray).
rlm@214 226
rlm@306 227 For convenience I've defined a few symbols for the more common
rlm@214 228 sensitivity values.
rlm@214 229
rlm@214 230 #+name: sensitivity
rlm@214 231 #+begin_src clojure
rlm@317 232 (def sensitivity-presets
rlm@317 233 "Retinal sensitivity presets for sensors that extract one channel
rlm@317 234 (:red :blue :green) or average all channels (:all)"
rlm@214 235 {:all 0xFFFFFF
rlm@214 236 :red 0xFF0000
rlm@214 237 :blue 0x0000FF
rlm@317 238 :green 0x00FF00})
rlm@214 239 #+end_src
rlm@214 240
rlm@214 241 ** Metadata Processing
rlm@214 242
rlm@273 243 =retina-sensor-profile= extracts a map from the eye-node in the same
rlm@273 244 format as the example maps above. =eye-dimensions= finds the
rlm@219 245 dimensions of the smallest image required to contain all the retinal
rlm@214 246 sensor maps.
rlm@214 247
rlm@216 248 #+name: retina
rlm@214 249 #+begin_src clojure
rlm@214 250 (defn retina-sensor-profile
rlm@214 251 "Return a map of pixel sensitivity numbers to BufferedImages
rlm@214 252 describing the distribution of light-sensitive components of this
rlm@214 253 eye. :red, :green, :blue, :gray are already defined as extracting
rlm@214 254 the red, green, blue, and average components respectively."
rlm@214 255 [#^Spatial eye]
rlm@214 256 (if-let [eye-map (meta-data eye "eye")]
rlm@214 257 (map-vals
rlm@214 258 load-image
rlm@214 259 (eval (read-string eye-map)))))
rlm@214 260
rlm@218 261 (defn eye-dimensions
rlm@218 262 "Returns [width, height] determined by the metadata of the eye."
rlm@214 263 [#^Spatial eye]
rlm@214 264 (let [dimensions
rlm@214 265 (map #(vector (.getWidth %) (.getHeight %))
rlm@214 266 (vals (retina-sensor-profile eye)))]
rlm@214 267 [(apply max (map first dimensions))
rlm@214 268 (apply max (map second dimensions))]))
rlm@214 269 #+end_src
rlm@214 270
ocsenave@265 271 * Importing and parsing descriptions of eyes.
rlm@214 272 First off, get the children of the "eyes" empty node to find all the
rlm@214 273 eyes the creature has.
rlm@216 274 #+name: eye-node
rlm@214 275 #+begin_src clojure
rlm@317 276 (def
rlm@317 277 ^{:doc "Return the children of the creature's \"eyes\" node."
rlm@317 278 :arglists '([creature])}
rlm@214 279 eyes
rlm@317 280 (sense-nodes "eyes"))
rlm@214 281 #+end_src
rlm@214 282
rlm@273 283 Then, add the camera created by =add-eye!= to the simulation by
rlm@215 284 creating a new viewport.
rlm@214 285
rlm@216 286 #+name: add-camera
rlm@213 287 #+begin_src clojure
rlm@338 288 (in-ns 'cortex.vision)
rlm@169 289 (defn add-camera!
rlm@169 290 "Add a camera to the world, calling continuation on every frame
rlm@34 291 produced."
rlm@167 292 [#^Application world camera continuation]
rlm@23 293 (let [width (.getWidth camera)
rlm@23 294 height (.getHeight camera)
rlm@23 295 render-manager (.getRenderManager world)
rlm@23 296 viewport (.createMainView render-manager "eye-view" camera)]
rlm@23 297 (doto viewport
rlm@23 298 (.setClearFlags true true true)
rlm@112 299 (.setBackgroundColor ColorRGBA/Black)
rlm@113 300 (.addProcessor (vision-pipeline continuation))
rlm@23 301 (.attachScene (.getRootNode world)))))
rlm@215 302 #+end_src
rlm@151 303
rlm@338 304 #+results: add-camera
rlm@338 305 : #'cortex.vision/add-camera!
rlm@338 306
rlm@151 307
rlm@218 308 The eye's continuation function should register the viewport with the
rlm@218 309 simulation the first time it is called, use the CPU to extract the
rlm@215 310 appropriate pixels from the rendered image and weight them by each
rlm@218 311 sensor's sensitivity. I have the option to do this processing in
rlm@218 312 native code for a slight gain in speed. I could also do it in the GPU
rlm@273 313 for a massive gain in speed. =vision-kernel= generates a list of
rlm@218 314 such continuation functions, one for each channel of the eye.
rlm@151 315
rlm@216 316 #+name: kernel
rlm@215 317 #+begin_src clojure
rlm@215 318 (in-ns 'cortex.vision)
rlm@151 319
rlm@215 320 (defrecord attached-viewport [vision-fn viewport-fn]
rlm@215 321 clojure.lang.IFn
rlm@215 322 (invoke [this world] (vision-fn world))
rlm@215 323 (applyTo [this args] (apply vision-fn args)))
rlm@151 324
rlm@216 325 (defn pixel-sense [sensitivity pixel]
rlm@216 326 (let [s-r (bit-shift-right (bit-and 0xFF0000 sensitivity) 16)
rlm@216 327 s-g (bit-shift-right (bit-and 0x00FF00 sensitivity) 8)
rlm@216 328 s-b (bit-and 0x0000FF sensitivity)
rlm@216 329
rlm@216 330 p-r (bit-shift-right (bit-and 0xFF0000 pixel) 16)
rlm@216 331 p-g (bit-shift-right (bit-and 0x00FF00 pixel) 8)
rlm@216 332 p-b (bit-and 0x0000FF pixel)
rlm@216 333
rlm@216 334 total-sensitivity (* 255 (+ s-r s-g s-b))]
rlm@216 335 (float (/ (+ (* s-r p-r)
rlm@216 336 (* s-g p-g)
rlm@216 337 (* s-b p-b))
rlm@216 338 total-sensitivity))))
rlm@216 339
rlm@215 340 (defn vision-kernel
rlm@171 341 "Returns a list of functions, each of which will return a color
rlm@171 342 channel's worth of visual information when called inside a running
rlm@171 343 simulation."
rlm@151 344 [#^Node creature #^Spatial eye & {skip :skip :or {skip 0}}]
rlm@169 345 (let [retinal-map (retina-sensor-profile eye)
rlm@169 346 camera (add-eye! creature eye)
rlm@151 347 vision-image
rlm@151 348 (atom
rlm@151 349 (BufferedImage. (.getWidth camera)
rlm@151 350 (.getHeight camera)
rlm@170 351 BufferedImage/TYPE_BYTE_BINARY))
rlm@170 352 register-eye!
rlm@170 353 (runonce
rlm@170 354 (fn [world]
rlm@170 355 (add-camera!
rlm@170 356 world camera
rlm@170 357 (let [counter (atom 0)]
rlm@170 358 (fn [r fb bb bi]
rlm@170 359 (if (zero? (rem (swap! counter inc) (inc skip)))
rlm@170 360 (reset! vision-image
rlm@170 361 (BufferedImage! r fb bb bi))))))))]
rlm@151 362 (vec
rlm@151 363 (map
rlm@151 364 (fn [[key image]]
rlm@151 365 (let [whites (white-coordinates image)
rlm@151 366 topology (vec (collapse whites))
rlm@216 367 sensitivity (sensitivity-presets key key)]
rlm@215 368 (attached-viewport.
rlm@215 369 (fn [world]
rlm@215 370 (register-eye! world)
rlm@215 371 (vector
rlm@215 372 topology
rlm@215 373 (vec
rlm@215 374 (for [[x y] whites]
rlm@216 375 (pixel-sense
rlm@216 376 sensitivity
rlm@216 377 (.getRGB @vision-image x y))))))
rlm@215 378 register-eye!)))
rlm@215 379 retinal-map))))
rlm@151 380
rlm@215 381 (defn gen-fix-display
rlm@215 382 "Create a function to call to restore a simulation's display when it
rlm@215 383 is disrupted by a Viewport."
rlm@215 384 []
rlm@215 385 (runonce
rlm@215 386 (fn [world]
rlm@215 387 (add-camera! world (.getCamera world) no-op))))
rlm@215 388 #+end_src
rlm@170 389
rlm@273 390 Note that since each of the functions generated by =vision-kernel=
rlm@273 391 shares the same =register-eye!= function, the eye will be registered
rlm@215 392 only once the first time any of the functions from the list returned
rlm@273 393 by =vision-kernel= is called. Each of the functions returned by
rlm@273 394 =vision-kernel= also allows access to the =Viewport= through which
rlm@306 395 it receives images.
rlm@215 396
rlm@306 397 The in-game display can be disrupted by all the ViewPorts that the
rlm@306 398 functions generated by =vision-kernel= add. This doesn't affect the
rlm@215 399 simulation or the simulated senses, but can be annoying.
rlm@273 400 =gen-fix-display= restores the in-simulation display.
rlm@215 401
ocsenave@265 402 ** The =vision!= function creates sensory probes.
rlm@215 403
rlm@218 404 All the hard work has been done; all that remains is to apply
rlm@273 405 =vision-kernel= to each eye in the creature and gather the results
rlm@215 406 into one list of functions.
rlm@215 407
rlm@216 408 #+name: main
rlm@215 409 #+begin_src clojure
rlm@170 410 (defn vision!
rlm@348 411 "Returns a list of functions, each of which returns visual sensory
rlm@348 412 data when called inside a running simulation."
rlm@151 413 [#^Node creature & {skip :skip :or {skip 0}}]
rlm@151 414 (reduce
rlm@170 415 concat
rlm@167 416 (for [eye (eyes creature)]
rlm@215 417 (vision-kernel creature eye))))
rlm@215 418 #+end_src
rlm@151 419
ocsenave@265 420 ** Displaying visual data for debugging.
ocsenave@265 421 # Visualization of Vision. Maybe less alliteration would be better.
rlm@215 422 It's vital to have a visual representation for each sense. Here I use
rlm@273 423 =view-sense= to construct a function that will create a display for
rlm@215 424 visual data.
rlm@215 425
rlm@216 426 #+name: display
rlm@215 427 #+begin_src clojure
rlm@216 428 (in-ns 'cortex.vision)
rlm@216 429
rlm@189 430 (defn view-vision
rlm@189 431 "Creates a function which accepts a list of visual sensor-data and
rlm@189 432 displays each element of the list to the screen."
rlm@189 433 []
rlm@188 434 (view-sense
rlm@188 435 (fn
rlm@188 436 [[coords sensor-data]]
rlm@188 437 (let [image (points->image coords)]
rlm@188 438 (dorun
rlm@188 439 (for [i (range (count coords))]
rlm@188 440 (.setRGB image ((coords i) 0) ((coords i) 1)
rlm@216 441 (gray (int (* 255 (sensor-data i)))))))
rlm@189 442 image))))
rlm@34 443 #+end_src
rlm@23 444
ocsenave@264 445 * Demonstrations
ocsenave@264 446 ** Demonstrating the vision pipeline.
rlm@23 447
rlm@215 448 This is a basic test for the vision system. It only tests the
ocsenave@264 449 vision-pipeline and does not deal with loading eyes from a blender
rlm@215 450 file. The code creates two videos of the same rotating cube from
rlm@215 451 different angles.
rlm@23 452
rlm@215 453 #+name: test-1
rlm@23 454 #+begin_src clojure
rlm@215 455 (in-ns 'cortex.test.vision)
rlm@23 456
rlm@219 457 (defn test-pipeline
rlm@69 458 "Testing vision:
rlm@69 459 Tests the vision system by creating two views of the same rotating
rlm@69 460 object from different angles and displaying both of those views in
rlm@69 461 JFrames.
rlm@69 462
rlm@69 463 You should see a rotating cube, and two windows,
rlm@69 464 each displaying a different view of the cube."
rlm@283 465 ([] (test-pipeline false))
rlm@283 466 ([record?]
rlm@283 467 (let [candy
rlm@283 468 (box 1 1 1 :physical? false :color ColorRGBA/Blue)]
rlm@283 469 (world
rlm@283 470 (doto (Node.)
rlm@283 471 (.attachChild candy))
rlm@283 472 {}
rlm@283 473 (fn [world]
rlm@283 474 (let [cam (.clone (.getCamera world))
rlm@283 475 width (.getWidth cam)
rlm@283 476 height (.getHeight cam)]
rlm@283 477 (add-camera! world cam
rlm@283 478 (comp
rlm@283 479 (view-image
rlm@283 480 (if record?
rlm@283 481 (File. "/home/r/proj/cortex/render/vision/1")))
rlm@283 482 BufferedImage!))
rlm@283 483 (add-camera! world
rlm@283 484 (doto (.clone cam)
rlm@283 485 (.setLocation (Vector3f. -10 0 0))
rlm@283 486 (.lookAt Vector3f/ZERO Vector3f/UNIT_Y))
rlm@283 487 (comp
rlm@283 488 (view-image
rlm@283 489 (if record?
rlm@283 490 (File. "/home/r/proj/cortex/render/vision/2")))
rlm@283 491 BufferedImage!))
rlm@341 492 (let [timer (IsoTimer. 60)]
rlm@340 493 (.setTimer world timer)
rlm@340 494 (display-dilated-time world timer))
rlm@283 495 ;; This is here to restore the main view
rlm@340 496 ;; after the other views have completed processing
rlm@283 497 (add-camera! world (.getCamera world) no-op)))
rlm@283 498 (fn [world tpf]
rlm@283 499 (.rotate candy (* tpf 0.2) 0 0))))))
rlm@23 500 #+end_src
rlm@23 501
rlm@340 502 #+results: test-1
rlm@340 503 : #'cortex.test.vision/test-pipeline
rlm@340 504
rlm@215 505 #+begin_html
rlm@215 506 <div class="figure">
rlm@215 507 <video controls="controls" width="755">
rlm@215 508 <source src="../video/spinning-cube.ogg" type="video/ogg"
rlm@215 509 preload="none" poster="../images/aurellem-1280x480.png" />
rlm@215 510 </video>
rlm@309 511 <br> <a href="http://youtu.be/r5Bn2aG7MO0"> YouTube </a>
rlm@215 512 <p>A rotating cube viewed from two different perspectives.</p>
rlm@215 513 </div>
rlm@215 514 #+end_html
rlm@215 515
rlm@215 516 Creating multiple eyes like this can be used for stereoscopic vision
rlm@215 517 simulation in a single creature or for simulating multiple creatures,
rlm@215 518 each with their own sense of vision.
ocsenave@264 519 ** Demonstrating eye import and parsing.
rlm@215 520
rlm@218 521 To the worm from the last post, I add a new node that describes its
rlm@215 522 eyes.
rlm@215 523
rlm@215 524 #+attr_html: width=755
rlm@215 525 #+caption: The worm with newly added empty nodes describing a single eye.
rlm@215 526 [[../images/worm-with-eye.png]]
rlm@215 527
rlm@215 528 The node highlighted in yellow is the root level "eyes" node. It has
rlm@218 529 a single child, highlighted in orange, which describes a single
rlm@218 530 eye. This is the "eye" node. It is placed so that the worm will have
rlm@218 531 an eye located in the center of the flat portion of its lower
rlm@218 532 hemispherical section.
rlm@218 533
rlm@218 534 The two nodes which are not highlighted describe the single joint of
rlm@218 535 the worm.
rlm@215 536
rlm@215 537 The metadata of the eye-node is:
rlm@215 538
rlm@215 539 #+begin_src clojure :results verbatim :exports both
rlm@215 540 (cortex.sense/meta-data
rlm@218 541 (.getChild (.getChild (cortex.test.body/worm) "eyes") "eye") "eye")
rlm@215 542 #+end_src
rlm@215 543
rlm@215 544 #+results:
rlm@215 545 : "(let [retina \"Models/test-creature/retina-small.png\"]
rlm@215 546 : {:all retina :red retina :green retina :blue retina})"
rlm@215 547
rlm@215 548 This is the approximation to the human eye described earlier.
rlm@215 549
rlm@216 550 #+name: test-2
rlm@215 551 #+begin_src clojure
rlm@215 552 (in-ns 'cortex.test.vision)
rlm@215 553
rlm@216 554 (defn change-color [obj color]
rlm@321 555 ;;(println-repl obj)
rlm@216 556 (if obj
rlm@216 557 (.setColor (.getMaterial obj) "Color" color)))
rlm@216 558
rlm@216 559 (defn colored-cannon-ball [color]
rlm@216 560 (comp #(change-color % color)
rlm@216 561 (fire-cannon-ball)))
rlm@215 562
rlm@338 563 (defn gen-worm
rlm@338 564 "create a creature acceptable for testing as a replacement for the
rlm@338 565 worm."
rlm@338 566 []
rlm@338 567 (nodify
rlm@338 568 "worm"
rlm@338 569 [(nodify
rlm@338 570 "eyes"
rlm@338 571 [(doto
rlm@338 572 (Node. "eye1")
rlm@338 573 (.setLocalTranslation (Vector3f. 0 -1.1 0))
rlm@338 574 (.setUserData
rlm@338 575
rlm@338 576 "eye"
rlm@338 577 "(let [retina
rlm@338 578 \"Models/test-creature/retina-small.png\"]
rlm@338 579 {:all retina :red retina
rlm@338 580 :green retina :blue retina})"))])
rlm@338 581 (box
rlm@338 582 0.2 0.2 0.2
rlm@338 583 :name "worm-segment"
rlm@338 584 :position (Vector3f. 0 0 0)
rlm@338 585 :color ColorRGBA/Orange)]))
rlm@338 586
rlm@338 587
rlm@338 588
rlm@283 589 (defn test-worm-vision
rlm@321 590 "Testing vision:
rlm@321 591 You should see the worm suspended in mid-air, looking down at a
rlm@321 592 table. There are four small displays, one each for red, green blue,
rlm@321 593 and gray channels. You can fire balls of various colors, and the
rlm@321 594 four channels should react accordingly.
rlm@321 595
rlm@321 596 Keys:
rlm@321 597 r : fire red-ball
rlm@321 598 b : fire blue-ball
rlm@321 599 g : fire green-ball
rlm@321 600 <space> : fire white ball"
rlm@338 601
rlm@283 602 ([] (test-worm-vision false))
rlm@283 603 ([record?]
rlm@283 604 (let [the-worm (doto (worm)(body!))
rlm@340 605 vision (vision! the-worm)
rlm@340 606 vision-display (view-vision)
rlm@340 607 fix-display (gen-fix-display)
rlm@283 608 me (sphere 0.5 :color ColorRGBA/Blue :physical? false)
rlm@283 609 x-axis
rlm@283 610 (box 1 0.01 0.01 :physical? false :color ColorRGBA/Red
rlm@283 611 :position (Vector3f. 0 -5 0))
rlm@283 612 y-axis
rlm@283 613 (box 0.01 1 0.01 :physical? false :color ColorRGBA/Green
rlm@283 614 :position (Vector3f. 0 -5 0))
rlm@283 615 z-axis
rlm@283 616 (box 0.01 0.01 1 :physical? false :color ColorRGBA/Blue
rlm@283 617 :position (Vector3f. 0 -5 0))
rlm@340 618
rlm@338 619 ]
rlm@215 620
rlm@335 621 (world
rlm@335 622 (nodify [(floor) the-worm x-axis y-axis z-axis me])
rlm@340 623 (merge standard-debug-controls
rlm@340 624 {"key-r" (colored-cannon-ball ColorRGBA/Red)
rlm@340 625 "key-b" (colored-cannon-ball ColorRGBA/Blue)
rlm@340 626 "key-g" (colored-cannon-ball ColorRGBA/Green)})
rlm@338 627
rlm@335 628 (fn [world]
rlm@340 629 (light-up-everything world)
rlm@340 630 (speed-up world)
rlm@341 631 (let [timer (IsoTimer. 60)]
rlm@340 632 (.setTimer world timer)
rlm@340 633 (display-dilated-time world timer))
rlm@340 634 ;; add a view from the worm's perspective
rlm@340 635 (if record?
rlm@340 636 (Capture/captureVideo
rlm@340 637 world
rlm@340 638 (File.
rlm@340 639 "/home/r/proj/cortex/render/worm-vision/main-view")))
rlm@340 640
rlm@340 641 (add-camera!
rlm@340 642 world
rlm@340 643 (add-eye! the-worm (first (eyes the-worm)))
rlm@340 644 (comp
rlm@340 645 (view-image
rlm@340 646 (if record?
rlm@340 647 (File.
rlm@340 648 "/home/r/proj/cortex/render/worm-vision/worm-view")))
rlm@340 649 BufferedImage!))
rlm@340 650
rlm@340 651 (set-gravity world Vector3f/ZERO)
rlm@340 652 (add-camera! world (.getCamera world) no-op))
rlm@340 653
rlm@340 654 (fn [world _]
rlm@340 655 (.setLocalTranslation me (.getLocation (.getCamera world)))
rlm@340 656 (vision-display
rlm@340 657 (map #(% world) vision)
rlm@338 658 (if record?
rlm@340 659 (File. "/home/r/proj/cortex/render/worm-vision")))
rlm@340 660 (fix-display world)
rlm@335 661 )))))
rlm@215 662 #+end_src
rlm@215 663
rlm@335 664 #+RESULTS: test-2
rlm@337 665 : #'cortex.test.vision/test-worm-vision
rlm@335 666
rlm@335 667
rlm@218 668 The world consists of the worm and a flat gray floor. I can shoot red,
rlm@218 669 green, blue and white cannonballs at the worm. The worm is initially
rlm@218 670 looking down at the floor, and there is no gravity. My perspective
rlm@218 671 (the Main View), the worm's perspective (Worm View) and the 4 sensor
rlm@218 672 channels that comprise the worm's eye are all saved frame-by-frame to
rlm@218 673 disk.
rlm@218 674
rlm@218 675 * Demonstration of Vision
rlm@218 676 #+begin_html
rlm@218 677 <div class="figure">
rlm@218 678 <video controls="controls" width="755">
rlm@218 679 <source src="../video/worm-vision.ogg" type="video/ogg"
rlm@218 680 preload="none" poster="../images/aurellem-1280x480.png" />
rlm@218 681 </video>
rlm@309 682 <br> <a href="http://youtu.be/J3H3iB_2NPQ"> YouTube </a>
rlm@218 683 <p>Simulated Vision in a Virtual Environment</p>
rlm@218 684 </div>
rlm@218 685 #+end_html
rlm@218 686
rlm@218 687 ** Generate the Worm Video from Frames
rlm@216 688 #+name: magick2
rlm@216 689 #+begin_src clojure
rlm@216 690 (ns cortex.video.magick2
rlm@216 691 (:import java.io.File)
rlm@316 692 (:use clojure.java.shell))
rlm@216 693
rlm@216 694 (defn images [path]
rlm@216 695 (sort (rest (file-seq (File. path)))))
rlm@216 696
rlm@216 697 (def base "/home/r/proj/cortex/render/worm-vision/")
rlm@216 698
rlm@216 699 (defn pics [file]
rlm@216 700 (images (str base file)))
rlm@216 701
rlm@216 702 (defn combine-images []
rlm@216 703 (let [main-view (pics "main-view")
rlm@216 704 worm-view (pics "worm-view")
rlm@216 705 blue (pics "0")
rlm@216 706 green (pics "1")
rlm@216 707 red (pics "2")
rlm@216 708 gray (pics "3")
rlm@216 709 blender (let [b-pics (pics "blender")]
rlm@216 710 (concat b-pics (repeat 9001 (last b-pics))))
rlm@216 711 background (repeat 9001 (File. (str base "background.png")))
rlm@216 712 targets (map
rlm@216 713 #(File. (str base "out/" (format "%07d.png" %)))
rlm@216 714 (range 0 (count main-view)))]
rlm@216 715 (dorun
rlm@216 716 (pmap
rlm@216 717 (comp
rlm@216 718 (fn [[background main-view worm-view red green blue gray blender target]]
rlm@216 719 (println target)
rlm@216 720 (sh "convert"
rlm@216 721 background
rlm@216 722 main-view "-geometry" "+18+17" "-composite"
rlm@216 723 worm-view "-geometry" "+677+17" "-composite"
rlm@216 724 green "-geometry" "+685+430" "-composite"
rlm@216 725 red "-geometry" "+788+430" "-composite"
rlm@216 726 blue "-geometry" "+894+430" "-composite"
rlm@216 727 gray "-geometry" "+1000+430" "-composite"
rlm@216 728 blender "-geometry" "+0+0" "-composite"
rlm@216 729 target))
rlm@216 730 (fn [& args] (map #(.getCanonicalPath %) args)))
rlm@216 731 background main-view worm-view red green blue gray blender targets))))
rlm@216 732 #+end_src
rlm@216 733
rlm@216 734 #+begin_src sh :results silent
rlm@216 735 cd /home/r/proj/cortex/render/worm-vision
rlm@216 736 ffmpeg -r 25 -b 9001k -i out/%07d.png -vcodec libtheora worm-vision.ogg
rlm@216 737 #+end_src
rlm@236 738
ocsenave@265 739 * Onward!
ocsenave@265 740 - As a neat bonus, this idea behind simulated vision also enables one
ocsenave@265 741 to [[../../cortex/html/capture-video.html][capture live video feeds from jMonkeyEngine]].
ocsenave@265 742 - Now that we have vision, it's time to tackle [[./hearing.org][hearing]].
ocsenave@265 743 #+appendix
ocsenave@265 744
rlm@215 745 * Headers
rlm@215 746
rlm@213 747 #+name: vision-header
rlm@213 748 #+begin_src clojure
rlm@213 749 (ns cortex.vision
rlm@213 750 "Simulate the sense of vision in jMonkeyEngine3. Enables multiple
rlm@213 751 eyes from different positions to observe the same world, and pass
rlm@306 752 the observed data to any arbitrary function. Automatically reads
rlm@216 753 eye-nodes from specially prepared blender files and instantiates
rlm@213 754 them in the world as actual eyes."
rlm@213 755 {:author "Robert McIntyre"}
rlm@213 756 (:use (cortex world sense util))
rlm@213 757 (:import com.jme3.post.SceneProcessor)
rlm@237 758 (:import (com.jme3.util BufferUtils Screenshots))
rlm@213 759 (:import java.nio.ByteBuffer)
rlm@213 760 (:import java.awt.image.BufferedImage)
rlm@213 761 (:import (com.jme3.renderer ViewPort Camera))
rlm@216 762 (:import (com.jme3.math ColorRGBA Vector3f Matrix3f))
rlm@213 763 (:import com.jme3.renderer.Renderer)
rlm@213 764 (:import com.jme3.app.Application)
rlm@213 765 (:import com.jme3.texture.FrameBuffer)
rlm@213 766 (:import (com.jme3.scene Node Spatial)))
rlm@213 767 #+end_src
rlm@112 768
rlm@215 769 #+name: test-header
rlm@215 770 #+begin_src clojure
rlm@215 771 (ns cortex.test.vision
rlm@215 772 (:use (cortex world sense util body vision))
rlm@215 773 (:use cortex.test.body)
rlm@215 774 (:import java.awt.image.BufferedImage)
rlm@215 775 (:import javax.swing.JPanel)
rlm@215 776 (:import javax.swing.SwingUtilities)
rlm@215 777 (:import java.awt.Dimension)
rlm@215 778 (:import javax.swing.JFrame)
rlm@215 779 (:import com.jme3.math.ColorRGBA)
rlm@215 780 (:import com.jme3.scene.Node)
rlm@215 781 (:import com.jme3.math.Vector3f)
rlm@216 782 (:import java.io.File)
rlm@341 783 (:import (com.aurellem.capture Capture RatchetTimer IsoTimer)))
rlm@215 784 #+end_src
rlm@341 785
rlm@341 786 #+results: test-header
rlm@341 787 : com.aurellem.capture.IsoTimer
rlm@341 788
rlm@216 789 * Source Listing
rlm@216 790 - [[../src/cortex/vision.clj][cortex.vision]]
rlm@216 791 - [[../src/cortex/test/vision.clj][cortex.test.vision]]
rlm@216 792 - [[../src/cortex/video/magick2.clj][cortex.video.magick2]]
rlm@216 793 - [[../assets/Models/subtitles/worm-vision-subtitles.blend][worm-vision-subtitles.blend]]
rlm@216 794 #+html: <ul> <li> <a href="../org/sense.org">This org file</a> </li> </ul>
rlm@216 795 - [[http://hg.bortreb.com ][source-repository]]
rlm@216 796
rlm@35 797
rlm@273 798 * Next
rlm@273 799 I find some [[./hearing.org][ears]] for the creature while exploring the guts of
rlm@273 800 jMonkeyEngine's sound system.
rlm@24 801
rlm@212 802 * COMMENT Generate Source
rlm@34 803 #+begin_src clojure :tangle ../src/cortex/vision.clj
rlm@216 804 <<vision-header>>
rlm@216 805 <<pipeline-1>>
rlm@216 806 <<pipeline-2>>
rlm@216 807 <<retina>>
rlm@216 808 <<add-eye>>
rlm@216 809 <<sensitivity>>
rlm@216 810 <<eye-node>>
rlm@216 811 <<add-camera>>
rlm@216 812 <<kernel>>
rlm@216 813 <<main>>
rlm@216 814 <<display>>
rlm@24 815 #+end_src
rlm@24 816
rlm@68 817 #+begin_src clojure :tangle ../src/cortex/test/vision.clj
rlm@215 818 <<test-header>>
rlm@215 819 <<test-1>>
rlm@216 820 <<test-2>>
rlm@24 821 #+end_src
rlm@216 822
rlm@216 823 #+begin_src clojure :tangle ../src/cortex/video/magick2.clj
rlm@216 824 <<magick2>>
rlm@216 825 #+end_src