view org/vision.org @ 548:0b891e0dd809

version 0.2 of thesis complete.
author Robert McIntyre <rlm@mit.edu>
date Thu, 01 May 2014 23:41:41 -0400
parents 3401053124b0
children 819968c8a391
line wrap: on
line source
1 #+title: Simulated Sense of Sight
2 #+author: Robert McIntyre
3 #+email: rlm@mit.edu
4 #+description: Simulated sight for AI research using JMonkeyEngine3 and clojure
5 #+keywords: computer vision, jMonkeyEngine3, clojure
6 #+SETUPFILE: ../../aurellem/org/setup.org
7 #+INCLUDE: ../../aurellem/org/level-0.org
8 #+babel: :mkdirp yes :noweb yes :exports both
10 * JMonkeyEngine natively supports multiple views of the same world.
12 Vision is one of the most important senses for humans, so I need to
13 build a simulated sense of vision for my AI. I will do this with
14 simulated eyes. Each eye can be independently moved and should see its
15 own version of the world depending on where it is.
17 Making these simulated eyes a reality is simple because jMonkeyEngine
18 already contains extensive support for multiple views of the same 3D
19 simulated world. The reason jMonkeyEngine has this support is because
20 the support is necessary to create games with split-screen
21 views. Multiple views are also used to create efficient
22 pseudo-reflections by rendering the scene from a certain perspective
23 and then projecting it back onto a surface in the 3D world.
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.
26 [[../images/goldeneye-4-player.png]]
28 ** =ViewPorts=, =SceneProcessors=, and the =RenderManager=.
29 # =ViewPorts= are cameras; =RenderManger= takes snapshots each frame.
30 #* A Brief Description of jMonkeyEngine's Rendering Pipeline
32 jMonkeyEngine allows you to create a =ViewPort=, which represents a
33 view of the simulated world. You can create as many of these as you
34 want. Every frame, the =RenderManager= iterates through each
35 =ViewPort=, rendering the scene in the GPU. For each =ViewPort= there
36 is a =FrameBuffer= which represents the rendered image in the GPU.
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.
39 #+ATTR_HTML: width="400"
40 [[../images/diagram_rendermanager2.png]]
42 Each =ViewPort= can have any number of attached =SceneProcessor=
43 objects, which are called every time a new frame is rendered. A
44 =SceneProcessor= receives its =ViewPort's= =FrameBuffer= and can do
45 whatever it wants to the data. Often this consists of invoking GPU
46 specific operations on the rendered image. The =SceneProcessor= can
47 also copy the GPU image data to RAM and process it with the CPU.
49 ** From Views to Vision
50 # Appropriating Views for Vision.
52 Each eye in the simulated creature needs its own =ViewPort= so that
53 it can see the world from its own perspective. To this =ViewPort=, I
54 add a =SceneProcessor= that feeds the visual data to any arbitrary
55 continuation function for further processing. That continuation
56 function may perform both CPU and GPU operations on the data. To make
57 this easy for the continuation function, the =SceneProcessor=
58 maintains appropriately sized buffers in RAM to hold the data. It does
59 not do any copying from the GPU to the CPU itself because it is a slow
60 operation.
62 #+name: pipeline-1
63 #+begin_src clojure
64 (defn vision-pipeline
65 "Create a SceneProcessor object which wraps a vision processing
66 continuation function. The continuation is a function that takes
67 [#^Renderer r #^FrameBuffer fb #^ByteBuffer b #^BufferedImage bi],
68 each of which has already been appropriately sized."
69 [continuation]
70 (let [byte-buffer (atom nil)
71 renderer (atom nil)
72 image (atom nil)]
73 (proxy [SceneProcessor] []
74 (initialize
75 [renderManager viewPort]
76 (let [cam (.getCamera viewPort)
77 width (.getWidth cam)
78 height (.getHeight cam)]
79 (reset! renderer (.getRenderer renderManager))
80 (reset! byte-buffer
81 (BufferUtils/createByteBuffer
82 (* width height 4)))
83 (reset! image (BufferedImage.
84 width height
85 BufferedImage/TYPE_4BYTE_ABGR))))
86 (isInitialized [] (not (nil? @byte-buffer)))
87 (reshape [_ _ _])
88 (preFrame [_])
89 (postQueue [_])
90 (postFrame
91 [#^FrameBuffer fb]
92 (.clear @byte-buffer)
93 (continuation @renderer fb @byte-buffer @image))
94 (cleanup []))))
95 #+end_src
97 The continuation function given to =vision-pipeline= above will be
98 given a =Renderer= and three containers for image data. The
99 =FrameBuffer= references the GPU image data, but the pixel data can
100 not be used directly on the CPU. The =ByteBuffer= and =BufferedImage=
101 are initially "empty" but are sized to hold the data in the
102 =FrameBuffer=. I call transferring the GPU image data to the CPU
103 structures "mixing" the image data. I have provided three functions to
104 do this mixing.
106 #+name: pipeline-2
107 #+begin_src clojure
108 (defn frameBuffer->byteBuffer!
109 "Transfer the data in the graphics card (Renderer, FrameBuffer) to
110 the CPU (ByteBuffer)."
111 [#^Renderer r #^FrameBuffer fb #^ByteBuffer bb]
112 (.readFrameBuffer r fb bb) bb)
114 (defn byteBuffer->bufferedImage!
115 "Convert the C-style BGRA image data in the ByteBuffer bb to the AWT
116 style ABGR image data and place it in BufferedImage bi."
117 [#^ByteBuffer bb #^BufferedImage bi]
118 (Screenshots/convertScreenShot bb bi) bi)
120 (defn BufferedImage!
121 "Continuation which will grab the buffered image from the materials
122 provided by (vision-pipeline)."
123 [#^Renderer r #^FrameBuffer fb #^ByteBuffer bb #^BufferedImage bi]
124 (byteBuffer->bufferedImage!
125 (frameBuffer->byteBuffer! r fb bb) bi))
126 #+end_src
128 Note that it is possible to write vision processing algorithms
129 entirely in terms of =BufferedImage= inputs. Just compose that
130 =BufferedImage= algorithm with =BufferedImage!=. However, a vision
131 processing algorithm that is entirely hosted on the GPU does not have
132 to pay for this convenience.
134 * Optical sensor arrays are described with images and referenced with metadata
135 The vision pipeline described above handles the flow of rendered
136 images. Now, we need simulated eyes to serve as the source of these
137 images.
139 An eye is described in blender in the same way as a joint. They are
140 zero dimensional empty objects with no geometry whose local coordinate
141 system determines the orientation of the resulting eye. All eyes are
142 children of a parent node named "eyes" just as all joints have a
143 parent named "joints". An eye binds to the nearest physical object
144 with =bind-sense=.
146 #+name: add-eye
147 #+begin_src clojure
148 (in-ns 'cortex.vision)
150 (defn add-eye!
151 "Create a Camera centered on the current position of 'eye which
152 follows the closest physical node in 'creature. The camera will
153 point in the X direction and use the Z vector as up as determined
154 by the rotation of these vectors in blender coordinate space. Use
155 XZY rotation for the node in blender."
156 [#^Node creature #^Spatial eye]
157 (let [target (closest-node creature eye)
158 [cam-width cam-height]
159 ;;[640 480] ;; graphics card on laptop doesn't support
160 ;; arbitray dimensions.
161 (eye-dimensions eye)
162 cam (Camera. cam-width cam-height)
163 rot (.getWorldRotation eye)]
164 (.setLocation cam (.getWorldTranslation eye))
165 (.lookAtDirection
166 cam ; this part is not a mistake and
167 (.mult rot Vector3f/UNIT_X) ; is consistent with using Z in
168 (.mult rot Vector3f/UNIT_Y)) ; blender as the UP vector.
169 (.setFrustumPerspective
170 cam (float 45)
171 (float (/ (.getWidth cam) (.getHeight cam)))
172 (float 1)
173 (float 1000))
174 (bind-sense target cam) cam))
175 #+end_src
177 Here, the camera is created based on metadata on the eye-node and
178 attached to the nearest physical object with =bind-sense=
179 ** The Retina
181 An eye is a surface (the retina) which contains many discrete sensors
182 to detect light. These sensors can have different light-sensing
183 properties. In humans, each discrete sensor is sensitive to red, blue,
184 green, or gray. These different types of sensors can have different
185 spatial distributions along the retina. In humans, there is a fovea in
186 the center of the retina which has a very high density of color
187 sensors, and a blind spot which has no sensors at all. Sensor density
188 decreases in proportion to distance from the fovea.
190 I want to be able to model any retinal configuration, so my eye-nodes
191 in blender contain metadata pointing to images that describe the
192 precise position of the individual sensors using white pixels. The
193 meta-data also describes the precise sensitivity to light that the
194 sensors described in the image have. An eye can contain any number of
195 these images. For example, the metadata for an eye might look like
196 this:
198 #+begin_src clojure
199 {0xFF0000 "Models/test-creature/retina-small.png"}
200 #+end_src
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.
203 [[../assets/Models/test-creature/retina-small.png]]
205 Together, the number 0xFF0000 and the image image above describe the
206 placement of red-sensitive sensory elements.
208 Meta-data to very crudely approximate a human eye might be something
209 like this:
211 #+begin_src clojure
212 (let [retinal-profile "Models/test-creature/retina-small.png"]
213 {0xFF0000 retinal-profile
214 0x00FF00 retinal-profile
215 0x0000FF retinal-profile
216 0xFFFFFF retinal-profile})
217 #+end_src
219 The numbers that serve as keys in the map determine a sensor's
220 relative sensitivity to the channels red, green, and blue. These
221 sensitivity values are packed into an integer in the order =|_|R|G|B|=
222 in 8-bit fields. The RGB values of a pixel in the image are added
223 together with these sensitivities as linear weights. Therefore,
224 0xFF0000 means sensitive to red only while 0xFFFFFF means sensitive to
225 all colors equally (gray).
227 For convenience I've defined a few symbols for the more common
228 sensitivity values.
230 #+name: sensitivity
231 #+begin_src clojure
232 (def sensitivity-presets
233 "Retinal sensitivity presets for sensors that extract one channel
234 (:red :blue :green) or average all channels (:all)"
235 {:all 0xFFFFFF
236 :red 0xFF0000
237 :blue 0x0000FF
238 :green 0x00FF00})
239 #+end_src
241 ** Metadata Processing
243 =retina-sensor-profile= extracts a map from the eye-node in the same
244 format as the example maps above. =eye-dimensions= finds the
245 dimensions of the smallest image required to contain all the retinal
246 sensor maps.
248 #+name: retina
249 #+begin_src clojure
250 (defn retina-sensor-profile
251 "Return a map of pixel sensitivity numbers to BufferedImages
252 describing the distribution of light-sensitive components of this
253 eye. :red, :green, :blue, :gray are already defined as extracting
254 the red, green, blue, and average components respectively."
255 [#^Spatial eye]
256 (if-let [eye-map (meta-data eye "eye")]
257 (map-vals
258 load-image
259 (eval (read-string eye-map)))))
261 (defn eye-dimensions
262 "Returns [width, height] determined by the metadata of the eye."
263 [#^Spatial eye]
264 (let [dimensions
265 (map #(vector (.getWidth %) (.getHeight %))
266 (vals (retina-sensor-profile eye)))]
267 [(apply max (map first dimensions))
268 (apply max (map second dimensions))]))
269 #+end_src
271 * Importing and parsing descriptions of eyes.
272 First off, get the children of the "eyes" empty node to find all the
273 eyes the creature has.
274 #+name: eye-node
275 #+begin_src clojure
276 (def
277 ^{:doc "Return the children of the creature's \"eyes\" node."
278 :arglists '([creature])}
279 eyes
280 (sense-nodes "eyes"))
281 #+end_src
283 Then, add the camera created by =add-eye!= to the simulation by
284 creating a new viewport.
286 #+name: add-camera
287 #+begin_src clojure
288 (in-ns 'cortex.vision)
289 (defn add-camera!
290 "Add a camera to the world, calling continuation on every frame
291 produced."
292 [#^Application world camera continuation]
293 (let [width (.getWidth camera)
294 height (.getHeight camera)
295 render-manager (.getRenderManager world)
296 viewport (.createMainView render-manager "eye-view" camera)]
297 (doto viewport
298 (.setClearFlags true true true)
299 (.setBackgroundColor ColorRGBA/Black)
300 (.addProcessor (vision-pipeline continuation))
301 (.attachScene (.getRootNode world)))))
302 #+end_src
304 #+results: add-camera
305 : #'cortex.vision/add-camera!
308 The eye's continuation function should register the viewport with the
309 simulation the first time it is called, use the CPU to extract the
310 appropriate pixels from the rendered image and weight them by each
311 sensor's sensitivity. I have the option to do this processing in
312 native code for a slight gain in speed. I could also do it in the GPU
313 for a massive gain in speed. =vision-kernel= generates a list of
314 such continuation functions, one for each channel of the eye.
316 #+name: kernel
317 #+begin_src clojure
318 (in-ns 'cortex.vision)
320 (defrecord attached-viewport [vision-fn viewport-fn]
321 clojure.lang.IFn
322 (invoke [this world] (vision-fn world))
323 (applyTo [this args] (apply vision-fn args)))
325 (defn pixel-sense [sensitivity pixel]
326 (let [s-r (bit-shift-right (bit-and 0xFF0000 sensitivity) 16)
327 s-g (bit-shift-right (bit-and 0x00FF00 sensitivity) 8)
328 s-b (bit-and 0x0000FF sensitivity)
330 p-r (bit-shift-right (bit-and 0xFF0000 pixel) 16)
331 p-g (bit-shift-right (bit-and 0x00FF00 pixel) 8)
332 p-b (bit-and 0x0000FF pixel)
334 total-sensitivity (* 255 (+ s-r s-g s-b))]
335 (float (/ (+ (* s-r p-r)
336 (* s-g p-g)
337 (* s-b p-b))
338 total-sensitivity))))
340 (defn vision-kernel
341 "Returns a list of functions, each of which will return a color
342 channel's worth of visual information when called inside a running
343 simulation."
344 [#^Node creature #^Spatial eye & {skip :skip :or {skip 0}}]
345 (let [retinal-map (retina-sensor-profile eye)
346 camera (add-eye! creature eye)
347 vision-image
348 (atom
349 (BufferedImage. (.getWidth camera)
350 (.getHeight camera)
351 BufferedImage/TYPE_BYTE_BINARY))
352 register-eye!
353 (runonce
354 (fn [world]
355 (add-camera!
356 world camera
357 (let [counter (atom 0)]
358 (fn [r fb bb bi]
359 (if (zero? (rem (swap! counter inc) (inc skip)))
360 (reset! vision-image
361 (BufferedImage! r fb bb bi))))))))]
362 (vec
363 (map
364 (fn [[key image]]
365 (let [whites (white-coordinates image)
366 topology (vec (collapse whites))
367 sensitivity (sensitivity-presets key key)]
368 (attached-viewport.
369 (fn [world]
370 (register-eye! world)
371 (vector
372 topology
373 (vec
374 (for [[x y] whites]
375 (pixel-sense
376 sensitivity
377 (.getRGB @vision-image x y))))))
378 register-eye!)))
379 retinal-map))))
381 (defn gen-fix-display
382 "Create a function to call to restore a simulation's display when it
383 is disrupted by a Viewport."
384 []
385 (runonce
386 (fn [world]
387 (add-camera! world (.getCamera world) no-op))))
388 #+end_src
390 Note that since each of the functions generated by =vision-kernel=
391 shares the same =register-eye!= function, the eye will be registered
392 only once the first time any of the functions from the list returned
393 by =vision-kernel= is called. Each of the functions returned by
394 =vision-kernel= also allows access to the =Viewport= through which
395 it receives images.
397 The in-game display can be disrupted by all the ViewPorts that the
398 functions generated by =vision-kernel= add. This doesn't affect the
399 simulation or the simulated senses, but can be annoying.
400 =gen-fix-display= restores the in-simulation display.
402 ** The =vision!= function creates sensory probes.
404 All the hard work has been done; all that remains is to apply
405 =vision-kernel= to each eye in the creature and gather the results
406 into one list of functions.
408 #+name: main
409 #+begin_src clojure
410 (defn vision!
411 "Returns a list of functions, each of which returns visual sensory
412 data when called inside a running simulation."
413 [#^Node creature & {skip :skip :or {skip 0}}]
414 (reduce
415 concat
416 (for [eye (eyes creature)]
417 (vision-kernel creature eye))))
418 #+end_src
420 ** Displaying visual data for debugging.
421 # Visualization of Vision. Maybe less alliteration would be better.
422 It's vital to have a visual representation for each sense. Here I use
423 =view-sense= to construct a function that will create a display for
424 visual data.
426 #+name: display
427 #+begin_src clojure
428 (in-ns 'cortex.vision)
430 (defn view-vision
431 "Creates a function which accepts a list of visual sensor-data and
432 displays each element of the list to the screen."
433 []
434 (view-sense
435 (fn
436 [[coords sensor-data]]
437 (let [image (points->image coords)]
438 (dorun
439 (for [i (range (count coords))]
440 (.setRGB image ((coords i) 0) ((coords i) 1)
441 (gray (int (* 255 (sensor-data i)))))))
442 image))))
443 #+end_src
445 * Demonstrations
446 ** Demonstrating the vision pipeline.
448 This is a basic test for the vision system. It only tests the
449 vision-pipeline and does not deal with loading eyes from a blender
450 file. The code creates two videos of the same rotating cube from
451 different angles.
453 #+name: test-1
454 #+begin_src clojure
455 (in-ns 'cortex.test.vision)
457 (defn test-pipeline
458 "Testing vision:
459 Tests the vision system by creating two views of the same rotating
460 object from different angles and displaying both of those views in
461 JFrames.
463 You should see a rotating cube, and two windows,
464 each displaying a different view of the cube."
465 ([] (test-pipeline false))
466 ([record?]
467 (let [candy
468 (box 1 1 1 :physical? false :color ColorRGBA/Blue)]
469 (world
470 (doto (Node.)
471 (.attachChild candy))
472 {}
473 (fn [world]
474 (let [cam (.clone (.getCamera world))
475 width (.getWidth cam)
476 height (.getHeight cam)]
477 (add-camera! world cam
478 (comp
479 (view-image
480 (if record?
481 (File. "/home/r/proj/cortex/render/vision/1")))
482 BufferedImage!))
483 (add-camera! world
484 (doto (.clone cam)
485 (.setLocation (Vector3f. -10 0 0))
486 (.lookAt Vector3f/ZERO Vector3f/UNIT_Y))
487 (comp
488 (view-image
489 (if record?
490 (File. "/home/r/proj/cortex/render/vision/2")))
491 BufferedImage!))
492 (let [timer (IsoTimer. 60)]
493 (.setTimer world timer)
494 (display-dilated-time world timer))
495 ;; This is here to restore the main view
496 ;; after the other views have completed processing
497 (add-camera! world (.getCamera world) no-op)))
498 (fn [world tpf]
499 (.rotate candy (* tpf 0.2) 0 0))))))
500 #+end_src
502 #+results: test-1
503 : #'cortex.test.vision/test-pipeline
505 #+begin_html
506 <div class="figure">
507 <video controls="controls" width="755">
508 <source src="../video/spinning-cube.ogg" type="video/ogg"
509 preload="none" poster="../images/aurellem-1280x480.png" />
510 </video>
511 <br> <a href="http://youtu.be/r5Bn2aG7MO0"> YouTube </a>
512 <p>A rotating cube viewed from two different perspectives.</p>
513 </div>
514 #+end_html
516 Creating multiple eyes like this can be used for stereoscopic vision
517 simulation in a single creature or for simulating multiple creatures,
518 each with their own sense of vision.
519 ** Demonstrating eye import and parsing.
521 To the worm from the last post, I add a new node that describes its
522 eyes.
524 #+attr_html: width=755
525 #+caption: The worm with newly added empty nodes describing a single eye.
526 [[../images/worm-with-eye.png]]
528 The node highlighted in yellow is the root level "eyes" node. It has
529 a single child, highlighted in orange, which describes a single
530 eye. This is the "eye" node. It is placed so that the worm will have
531 an eye located in the center of the flat portion of its lower
532 hemispherical section.
534 The two nodes which are not highlighted describe the single joint of
535 the worm.
537 The metadata of the eye-node is:
539 #+begin_src clojure :results verbatim :exports both
540 (cortex.sense/meta-data
541 (.getChild (.getChild (cortex.test.body/worm) "eyes") "eye") "eye")
542 #+end_src
544 #+results:
545 : "(let [retina \"Models/test-creature/retina-small.png\"]
546 : {:all retina :red retina :green retina :blue retina})"
548 This is the approximation to the human eye described earlier.
550 #+name: test-2
551 #+begin_src clojure
552 (in-ns 'cortex.test.vision)
554 (defn change-color [obj color]
555 ;;(println-repl obj)
556 (if obj
557 (.setColor (.getMaterial obj) "Color" color)))
559 (defn colored-cannon-ball [color]
560 (comp #(change-color % color)
561 (fire-cannon-ball)))
563 (defn gen-worm
564 "create a creature acceptable for testing as a replacement for the
565 worm."
566 []
567 (nodify
568 "worm"
569 [(nodify
570 "eyes"
571 [(doto
572 (Node. "eye1")
573 (.setLocalTranslation (Vector3f. 0 -1.1 0))
574 (.setUserData
576 "eye"
577 "(let [retina
578 \"Models/test-creature/retina-small.png\"]
579 {:all retina :red retina
580 :green retina :blue retina})"))])
581 (box
582 0.2 0.2 0.2
583 :name "worm-segment"
584 :position (Vector3f. 0 0 0)
585 :color ColorRGBA/Orange)]))
589 (defn test-worm-vision
590 "Testing vision:
591 You should see the worm suspended in mid-air, looking down at a
592 table. There are four small displays, one each for red, green blue,
593 and gray channels. You can fire balls of various colors, and the
594 four channels should react accordingly.
596 Keys:
597 r : fire red-ball
598 b : fire blue-ball
599 g : fire green-ball
600 <space> : fire white ball"
602 ([] (test-worm-vision false))
603 ([record?]
604 (let [the-worm (doto (worm)(body!))
605 vision (vision! the-worm)
606 vision-display (view-vision)
607 fix-display (gen-fix-display)
608 me (sphere 0.5 :color ColorRGBA/Blue :physical? false)
609 x-axis
610 (box 1 0.01 0.01 :physical? false :color ColorRGBA/Red
611 :position (Vector3f. 0 -5 0))
612 y-axis
613 (box 0.01 1 0.01 :physical? false :color ColorRGBA/Green
614 :position (Vector3f. 0 -5 0))
615 z-axis
616 (box 0.01 0.01 1 :physical? false :color ColorRGBA/Blue
617 :position (Vector3f. 0 -5 0))
619 ]
621 (world
622 (nodify [(floor) the-worm x-axis y-axis z-axis me])
623 (merge standard-debug-controls
624 {"key-r" (colored-cannon-ball ColorRGBA/Red)
625 "key-b" (colored-cannon-ball ColorRGBA/Blue)
626 "key-g" (colored-cannon-ball ColorRGBA/Green)})
628 (fn [world]
629 (light-up-everything world)
630 (speed-up world)
631 (let [timer (IsoTimer. 60)]
632 (.setTimer world timer)
633 (display-dilated-time world timer))
634 ;; add a view from the worm's perspective
635 (if record?
636 (Capture/captureVideo
637 world
638 (File.
639 "/home/r/proj/cortex/render/worm-vision/main-view")))
641 (add-camera!
642 world
643 (add-eye! the-worm (first (eyes the-worm)))
644 (comp
645 (view-image
646 (if record?
647 (File.
648 "/home/r/proj/cortex/render/worm-vision/worm-view")))
649 BufferedImage!))
651 (set-gravity world Vector3f/ZERO)
652 (add-camera! world (.getCamera world) no-op))
654 (fn [world _]
655 (.setLocalTranslation me (.getLocation (.getCamera world)))
656 (vision-display
657 (map #(% world) vision)
658 (if record?
659 (File. "/home/r/proj/cortex/render/worm-vision")))
660 (fix-display world)
661 )))))
662 #+end_src
664 #+RESULTS: test-2
665 : #'cortex.test.vision/test-worm-vision
668 The world consists of the worm and a flat gray floor. I can shoot red,
669 green, blue and white cannonballs at the worm. The worm is initially
670 looking down at the floor, and there is no gravity. My perspective
671 (the Main View), the worm's perspective (Worm View) and the 4 sensor
672 channels that comprise the worm's eye are all saved frame-by-frame to
673 disk.
675 * Demonstration of Vision
676 #+begin_html
677 <div class="figure">
678 <video controls="controls" width="755">
679 <source src="../video/worm-vision.ogg" type="video/ogg"
680 preload="none" poster="../images/aurellem-1280x480.png" />
681 </video>
682 <br> <a href="http://youtu.be/J3H3iB_2NPQ"> YouTube </a>
683 <p>Simulated Vision in a Virtual Environment</p>
684 </div>
685 #+end_html
687 ** Generate the Worm Video from Frames
688 #+name: magick2
689 #+begin_src clojure
690 (ns cortex.video.magick2
691 (:import java.io.File)
692 (:use clojure.java.shell))
694 (defn images [path]
695 (sort (rest (file-seq (File. path)))))
697 (def base "/home/r/proj/cortex/render/worm-vision/")
699 (defn pics [file]
700 (images (str base file)))
702 (defn combine-images []
703 (let [main-view (pics "main-view")
704 worm-view (pics "worm-view")
705 blue (pics "0")
706 green (pics "1")
707 red (pics "2")
708 gray (pics "3")
709 blender (let [b-pics (pics "blender")]
710 (concat b-pics (repeat 9001 (last b-pics))))
711 background (repeat 9001 (File. (str base "background.png")))
712 targets (map
713 #(File. (str base "out/" (format "%07d.png" %)))
714 (range 0 (count main-view)))]
715 (dorun
716 (pmap
717 (comp
718 (fn [[background main-view worm-view red green blue gray blender target]]
719 (println target)
720 (sh "convert"
721 background
722 main-view "-geometry" "+18+17" "-composite"
723 worm-view "-geometry" "+677+17" "-composite"
724 green "-geometry" "+685+430" "-composite"
725 red "-geometry" "+788+430" "-composite"
726 blue "-geometry" "+894+430" "-composite"
727 gray "-geometry" "+1000+430" "-composite"
728 blender "-geometry" "+0+0" "-composite"
729 target))
730 (fn [& args] (map #(.getCanonicalPath %) args)))
731 background main-view worm-view red green blue gray blender targets))))
732 #+end_src
734 #+begin_src sh :results silent
735 cd /home/r/proj/cortex/render/worm-vision
736 ffmpeg -r 25 -b 9001k -i out/%07d.png -vcodec libtheora worm-vision.ogg
737 #+end_src
739 * Onward!
740 - As a neat bonus, this idea behind simulated vision also enables one
741 to [[../../cortex/html/capture-video.html][capture live video feeds from jMonkeyEngine]].
742 - Now that we have vision, it's time to tackle [[./hearing.org][hearing]].
743 #+appendix
745 * Headers
747 #+name: vision-header
748 #+begin_src clojure
749 (ns cortex.vision
750 "Simulate the sense of vision in jMonkeyEngine3. Enables multiple
751 eyes from different positions to observe the same world, and pass
752 the observed data to any arbitrary function. Automatically reads
753 eye-nodes from specially prepared blender files and instantiates
754 them in the world as actual eyes."
755 {:author "Robert McIntyre"}
756 (:use (cortex world sense util))
757 (:import com.jme3.post.SceneProcessor)
758 (:import (com.jme3.util BufferUtils Screenshots))
759 (:import java.nio.ByteBuffer)
760 (:import java.awt.image.BufferedImage)
761 (:import (com.jme3.renderer ViewPort Camera))
762 (:import (com.jme3.math ColorRGBA Vector3f Matrix3f))
763 (:import com.jme3.renderer.Renderer)
764 (:import com.jme3.app.Application)
765 (:import com.jme3.texture.FrameBuffer)
766 (:import (com.jme3.scene Node Spatial)))
767 #+end_src
769 #+name: test-header
770 #+begin_src clojure
771 (ns cortex.test.vision
772 (:use (cortex world sense util body vision))
773 (:use cortex.test.body)
774 (:import java.awt.image.BufferedImage)
775 (:import javax.swing.JPanel)
776 (:import javax.swing.SwingUtilities)
777 (:import java.awt.Dimension)
778 (:import javax.swing.JFrame)
779 (:import com.jme3.math.ColorRGBA)
780 (:import com.jme3.scene.Node)
781 (:import com.jme3.math.Vector3f)
782 (:import java.io.File)
783 (:import (com.aurellem.capture Capture RatchetTimer IsoTimer)))
784 #+end_src
786 #+results: test-header
787 : com.aurellem.capture.IsoTimer
789 * Source Listing
790 - [[../src/cortex/vision.clj][cortex.vision]]
791 - [[../src/cortex/test/vision.clj][cortex.test.vision]]
792 - [[../src/cortex/video/magick2.clj][cortex.video.magick2]]
793 - [[../assets/Models/subtitles/worm-vision-subtitles.blend][worm-vision-subtitles.blend]]
794 #+html: <ul> <li> <a href="../org/sense.org">This org file</a> </li> </ul>
795 - [[http://hg.bortreb.com ][source-repository]]
798 * Next
799 I find some [[./hearing.org][ears]] for the creature while exploring the guts of
800 jMonkeyEngine's sound system.
802 * COMMENT Generate Source
803 #+begin_src clojure :tangle ../src/cortex/vision.clj
804 <<vision-header>>
805 <<pipeline-1>>
806 <<pipeline-2>>
807 <<retina>>
808 <<add-eye>>
809 <<sensitivity>>
810 <<eye-node>>
811 <<add-camera>>
812 <<kernel>>
813 <<main>>
814 <<display>>
815 #+end_src
817 #+begin_src clojure :tangle ../src/cortex/test/vision.clj
818 <<test-header>>
819 <<test-1>>
820 <<test-2>>
821 #+end_src
823 #+begin_src clojure :tangle ../src/cortex/video/magick2.clj
824 <<magick2>>
825 #+end_src