-
Notifications
You must be signed in to change notification settings - Fork 48
/
Copy pathminimal__no_modules__track_one_face__performance_comparison.html
368 lines (234 loc) · 11.3 KB
/
minimal__no_modules__track_one_face__performance_comparison.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="utf-8">
<title>Beyond Reality Face SDK - BRFv5 - Face Tracking for Browser/JavaScript - Minimal Webcam Example (no modules)</title>
<style>
html, body { width: 100%; height: 100%; background-color: #ffffff; margin: 0; padding: 0; overflow: hidden; }
</style>
</head>
<body>
<!--
A little walk through the basics ...
We need at least a video for the camera stream and a canvas for drawing the camera stream (mirrored),
retrieving the image data and drawing the results.
In the modules version, we have two canvases and only update BRFv5 when a new image was received from
the camera stream. Here we only have one canvas and draw the video with each requestAnimationFrame callback.
This minimal example does not use modules, but plain JavaScript, so the '_no_modules.js' script places the
brfv5Module function right into the window namespace (the script has no exports).
You can set a URL param: fast=1 to quickly change from normal to fast execution.
-->
<video id="_webcam" style="display: none;" playsinline></video>
<canvas id="_imageData"></canvas>
<script src="./js/brfv5/brfv5_js_tk261120_v5.2.1_trial_no_modules.js"></script>
<script src="./js/ui/stats.no_modules.js"></script>
<script>
// Set the BRFv5 library name here, also set your own appId for reference.
const _libraryName = 'brfv5_js_tk261120_v5.2.1_trial.brfv5'
const _appId = 'brfv5.browser.minimal.nomodules' // (mandatory): 8 to 64 characters, a-z . 0-9 allowed
const brfv5 = {} // The library namespace.
// References to the video and canvas.
const _webcam = document.getElementById('_webcam')
const _imageData = document.getElementById('_imageData')
// Those variables will be retrieved from the stream and the library.
let _brfv5Manager = null
let _brfv5Config = null
let _width = 0
let _height = 0
const _stats = new Stats()
document.body.appendChild(_stats.dom)
const _times = [ 33 ]
// Settings to compare execution times via Stats:
const setNormal = {
modelType: '68l', // tracks 68 landmarks
numChunksToLoad: 8, // number of cascade chunks. 8 is max, 1 (only 3 cascades) is mininum
numTrackingPasses: 3, // max is 6, which is not recommended, min is 1
enableFreeRotation: true // true: tilt can go beyound 35 degrees, wasn't possible in v4
}
const setFaster = {
modelType: '42l', // tracks 42 landmarks
numChunksToLoad: 3, // number of cascade chunks. 8 is max, 1 (only 3 cascades) is mininum
numTrackingPasses: 3, // max is 6, which is not recommended, min is 1,
enableFreeRotation: false // true: tilt can go beyound 35 degrees, wasn't possible in v4
}
const urlParams = new URLSearchParams(window.location.search)
const settings = urlParams.get('fast') === '1' ? setFaster : setNormal
// loadBRFv5Model and openCamera are being done simultaneously thanks to Promises. Both call
// configureTracking which only gets executed once both Promises were successful. Once configured
// trackFaces will do the tracking work and draw the results.
const loadBRFv5Model = (modelName, numChunksToLoad, pathToModels = '', appId = null, onProgress = null) => {
console.log('loadBRFv5Model')
if(!modelName) { throw 'Please provide a modelName.' }
return new Promise((resolve, reject) => {
if(_brfv5Manager && _brfv5Config) {
resolve({ brfv5Manager: _brfv5Manager, brfv5Config: _brfv5Config })
} else {
try {
brfv5.appId = appId ? appId : _appId
brfv5.binaryLocation = pathToModels + _libraryName
brfv5.modelLocation = pathToModels + modelName + '_c'
brfv5.modelChunks = numChunksToLoad // 4, 6, 8
brfv5.binaryProgress = onProgress
brfv5.binaryError = (e) => { reject(e) }
brfv5.onInit = (brfv5Manager, brfv5Config) => {
_brfv5Manager = brfv5Manager
_brfv5Config = brfv5Config
resolve({ brfv5Manager: _brfv5Manager, brfv5Config: _brfv5Config })
}
brfv5Module(brfv5)
} catch(e) {
reject(e)
}
}
})
}
const openCamera = () => {
console.log('openCamera')
return new Promise((resolve, reject) => {
window.navigator.mediaDevices.getUserMedia({ video: { width: 1280, height: 720, frameRate: 30, facingMode: 'user'} })
.then((mediaStream) => {
_webcam.srcObject = mediaStream
_webcam.play().then(() => { resolve({ width: _webcam.videoWidth, height: _webcam.videoHeight }) }).catch((e) => { reject(e) })
}).catch((e) => { reject(e) })
})
}
const configureTracking = () => {
if(_brfv5Config !== null && _width > 0) {
// Camera stream and BRFv5 are ready. Now configure. Internal defaults are set for a 640x480 resolution.
// So the following isn't really necessary.
const brfv5Config = _brfv5Config
const imageWidth = _width
const imageHeight = _height
const inputSize = imageWidth > imageHeight ? imageHeight : imageWidth
// Setup image data dimensions
brfv5Config.imageConfig.inputWidth = imageWidth
brfv5Config.imageConfig.inputHeight = imageHeight
const sizeFactor = inputSize / 480.0
// Set face detection region of interest and parameters scaled to the image base size.
brfv5Config.faceDetectionConfig.regionOfInterest.setTo(0, 0, imageWidth, imageHeight)
brfv5Config.faceDetectionConfig.minFaceSize = 144 * sizeFactor
brfv5Config.faceDetectionConfig.maxFaceSize = 480 * sizeFactor
if(imageWidth < imageHeight) {
// Portrait mode: probably smartphone, faces tend to be closer to the camera, processing time is an issue,
// so save a bit of time and increase minFaceSize.
brfv5Config.faceDetectionConfig.minFaceSize = 240 * sizeFactor
}
// Set face tracking region of interest and parameters scaled to the image base size.
brfv5Config.faceTrackingConfig.regionOfInterest.setTo(0, 0, imageWidth, imageHeight)
brfv5Config.faceTrackingConfig.minFaceScaleStart = 50.0 * sizeFactor
brfv5Config.faceTrackingConfig.maxFaceScaleStart = 320.0 * sizeFactor
brfv5Config.faceTrackingConfig.minFaceScaleReset = 35.0 * sizeFactor
brfv5Config.faceTrackingConfig.maxFaceScaleReset = 420.0 * sizeFactor
brfv5Config.faceTrackingConfig.confidenceThresholdReset = 0.001
brfv5Config.faceTrackingConfig.enableStabilizer = true
brfv5Config.faceTrackingConfig.maxRotationXReset = 35.0
brfv5Config.faceTrackingConfig.maxRotationYReset = 45.0
brfv5Config.faceTrackingConfig.maxRotationZReset = 34.0
brfv5Config.faceTrackingConfig.numTrackingPasses = settings.numTrackingPasses
brfv5Config.faceTrackingConfig.enableFreeRotation = settings.enableFreeRotation
brfv5Config.faceTrackingConfig.maxRotationZReset = settings.enableFreeRotation ? 999.0 : 34.0
brfv5Config.faceTrackingConfig.numFacesToTrack = 1
brfv5Config.enableFaceTracking = true
console.log('configureTracking:', _brfv5Config)
_brfv5Manager.configure(_brfv5Config)
trackFaces()
}
}
const trackFaces = () => {
if(!_brfv5Manager || !_brfv5Config || !_imageData) { return }
const timeStart = Date.now()
const ctx = _imageData.getContext('2d')
ctx.setTransform(-1.0, 0, 0, 1, _width, 0) // A virtual mirror should be... mirrored
ctx.drawImage(_webcam, 0, 0, _width, _height)
ctx.setTransform(1.0, 0, 0, 1, 0, 0) // unmirror to draw the results
_brfv5Manager.update(ctx.getImageData(0, 0, _width, _height))
let doDrawFaceDetection = !_brfv5Config.enableFaceTracking
if(_brfv5Config.enableFaceTracking) {
const sizeFactor = Math.min(_width, _height) / 480.0
const faces = _brfv5Manager.getFaces()
for(let i = 0; i < faces.length; i++) {
const face = faces[i]
if(face.state === brfv5.BRFv5State.FACE_TRACKING) {
drawRect(ctx, _brfv5Config.faceTrackingConfig.regionOfInterest, '#00a0ff', 2.0)
drawCircles(ctx, face.landmarks, '#00a0ff', 2.0 * sizeFactor)
drawRect(ctx, face.bounds, '#ffffff', 1.0)
} else {
doDrawFaceDetection = true
}
}
}
if(doDrawFaceDetection) {
// Only draw face detection results, if face detection was performed.
drawRect( ctx, _brfv5Config.faceDetectionConfig.regionOfInterest, '#ffffff', 2.0)
drawRects(ctx, _brfv5Manager.getDetectedRects(), '#00a0ff', 1.0)
drawRects(ctx, _brfv5Manager.getMergedRects(), '#ffffff', 3.0)
}
const timeEnd = Date.now()
_times.push(timeEnd - timeStart)
while(_times.length > 30) { _times.shift() }
let _averageTime = 0
for(let i = 0; i < _times.length; i++) { _averageTime += _times[i]; }
_averageTime /= _times.length
_stats.setMS(_averageTime)
requestAnimationFrame(trackFaces)
}
openCamera().then(({ width, height }) => {
console.log('openCamera: done: ' + width + 'x' + height)
_width = width
_height = height
_imageData.width = _width
_imageData.height = _height
configureTracking()
}).catch((e) => { if(e) { console.error('Camera failed: ', e) } })
loadBRFv5Model(settings.modelType, settings.numChunksToLoad, './js/brfv5/models/', _appId,
(progress) => { console.log(progress) }).then(({ brfv5Manager, brfv5Config }) => {
console.log('loadBRFv5Model: done')
_brfv5Manager = brfv5Manager
_brfv5Config = brfv5Config
configureTracking()
}).catch((e) => { console.error('BRFv5 failed: ', e) })
const drawCircles = (ctx, array, color, radius) => {
ctx.strokeStyle = null
ctx.fillStyle = getColor(color, 1.0)
let _radius = radius || 2.0
for(let i = 0; i < array.length; ++i) {
ctx.beginPath()
ctx.arc(array[i].x, array[i].y, _radius, 0, 2 * Math.PI)
ctx.fill()
}
}
const drawRect = (ctx, rect, color, lineWidth) => {
ctx.strokeStyle = getColor(color, 1.0)
ctx.fillStyle = null
ctx.lineWidth = lineWidth || 1.0
ctx.beginPath()
ctx.rect(rect.x, rect.y, rect.width, rect.height)
ctx.stroke()
}
const drawRects = (ctx, rects, color, lineWidth) => {
ctx.strokeStyle = getColor(color, 1.0)
ctx.fillStyle = null
ctx.lineWidth = lineWidth || 1.0
for(let i = 0; i < rects.length; ++i) {
let rect = rects[i]
ctx.beginPath()
ctx.rect(rect.x, rect.y, rect.width, rect.height)
ctx.stroke()
}
}
const getColor = (color, alpha) => {
const colorStr = color + ''
if(colorStr.startsWith('rgb')) {
return color
}
if(colorStr.startsWith('#')) {
color = parseInt('0x' + colorStr.substr(1))
}
return 'rgb(' +
(((color >> 16) & 0xff).toString(10)) + ', ' +
(((color >> 8) & 0xff).toString(10)) + ', ' +
(((color) & 0xff).toString(10)) + ', ' + alpha +')'
}
</script>
</body>
</html>