-
Notifications
You must be signed in to change notification settings - Fork 8
/
camera.js
481 lines (423 loc) · 15.1 KB
/
camera.js
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
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
/**
* @license
* Copyright 2019 Google LLC. All Rights Reserved.
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* https://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
* =============================================================================
*/
import * as posenet from '@tensorflow-models/posenet';
import dat from 'dat.gui';
import Stats from 'stats.js';
import {drawBoundingBox, drawKeypoints, drawSkeleton, isMobile, toggleLoadingUI, global_zero} from './demo_util';
const videoWidth = 600;
const videoHeight = 500;
var rep_count = 0;
var wko_started = 0;
var done = 0;
var timer = 0;
var t0 = 0;
var count_down = 0;
var act_dur = 0;
var last_rep = 0;
const stats = new Stats();
/**
* Loads a the camera to be used in the demo
*
*/
async function setupCamera() {
if (!navigator.mediaDevices || !navigator.mediaDevices.getUserMedia) {
throw new Error(
'Browser API navigator.mediaDevices.getUserMedia not available');
}
const video = document.getElementById('video');
video.width = videoWidth;
video.height = videoHeight;
const mobile = isMobile();
const stream = await navigator.mediaDevices.getUserMedia({
'audio': false,
'video': {
facingMode: 'user',
width: mobile ? undefined : videoWidth,
height: mobile ? undefined : videoHeight,
},
});
video.srcObject = stream;
return new Promise((resolve) => {
video.onloadedmetadata = () => {
resolve(video);
};
});
}
async function loadVideo() {
const video = await setupCamera();
video.play();
return video;
}
const defaultQuantBytes = 2;
const defaultMobileNetMultiplier = isMobile() ? 0.50 : 0.75;
const defaultMobileNetStride = 16;
const defaultMobileNetInputResolution = 500;
const guiState = {
algorithm: 'multi-pose',
input: {
architecture: 'MobileNetV1',
outputStride: defaultMobileNetStride,
inputResolution: defaultMobileNetInputResolution,
multiplier: defaultMobileNetMultiplier,
quantBytes: defaultQuantBytes
},
activity: {
Activity: 'Weight Lifting',
Repetitions: 10,
Duration_min: 60
},
singlePoseDetection: {
minPoseConfidence: 0.1,
minPartConfidence: 0.5,
},
multiPoseDetection: {
maxPoseDetections: 5,
minPoseConfidence: 0.15,
minPartConfidence: 0.1,
nmsRadius: 30.0,
},
output: {
showVideo: true,
showSkeleton: true,
showPoints: true,
showBoundingBox: false,
},
net: null,
};
/**
* Sets up dat.gui controller on the top-right of the window
*/
function setupGui(cameras, net) {
guiState.net = net;
if (cameras.length > 0) {
guiState.camera = cameras[0].deviceId;
}
const gui = new dat.GUI({width: 300});
let architectureController = null;
// guiState[tryResNetButtonName] = function() {
// architectureController.setValue('ResNet50')
// };
//gui.add(guiState, tryResNetButtonName).name(tryResNetButtonText);
//updateTryResNetButtonDatGuiCss();
// The single-pose algorithm is faster and simpler but requires only one
// person to be in the frame or results will be innaccurate. Multi-pose works
// for more than 1 person
const algorithmController =
gui.add(guiState, 'algorithm', ['single-pose', 'multi-pose']);
// The input parameters have the most effect on accuracy and speed of the
// network
let activityController = null;
let activity = gui.addFolder('Workout');
activityController =
activity.add(guiState.activity, 'Activity', ['Weight Lifting', 'Jumping jack']);
guiState.atividade = guiState.activity.activity;
activity.add(guiState.activity, 'Repetitions', 0, 50);
activity.add(guiState.activity, 'Duration_min', 0, 180).name('Duration (sec)');
var obj = { add:function(){
console.log("clicked");
rep_count = 0;
count_down = 0;
wko_started = 1
done = 0
global_zero();
}};
activity.add(obj,'add').name('Start Workout');
activity.open();
let single = gui.addFolder('Single Pose Detection');
single.add(guiState.singlePoseDetection, 'minPoseConfidence', 0.0, 1.0);
single.add(guiState.singlePoseDetection, 'minPartConfidence', 0.0, 1.0);
let multi = gui.addFolder('Multi Pose Detection');
multi.add(guiState.multiPoseDetection, 'maxPoseDetections')
.min(1)
.max(20)
.step(1);
multi.add(guiState.multiPoseDetection, 'minPoseConfidence', 0.0, 1.0);
multi.add(guiState.multiPoseDetection, 'minPartConfidence', 0.0, 1.0);
// nms Radius: controls the minimum distance between poses that are returned
// defaults to 20, which is probably fine for most use cases
multi.add(guiState.multiPoseDetection, 'nmsRadius').min(0.0).max(40.0);
multi.open();
let output = gui.addFolder('Output');
output.add(guiState.output, 'showVideo');
output.add(guiState.output, 'showSkeleton');
output.add(guiState.output, 'showPoints');
output.add(guiState.output, 'showBoundingBox');
output.open();
algorithmController.onChange(function(value) {
switch (guiState.algorithm) {
case 'single-pose':
multi.close();
single.open();
break;
case 'multi-pose':
single.close();
multi.open();
break;
}
});
}
/**
* Sets up a frames per second panel on the top-left of the window
*/
function setupFPS() {
stats.showPanel(0); // 0: fps, 1: ms, 2: mb, 3+: custom
document.getElementById('main').appendChild(stats.dom);
}
/**
* Feeds an image to posenet to estimate poses - this is where the magic
* happens. This function loops with a requestAnimationFrame method.
*/
function detectPoseInRealTime(video, net) {
const canvas = document.getElementById('output');
const ctx = canvas.getContext('2d');
// since images are being fed from a webcam, we want to feed in the
// original image and then just flip the keypoints' x coordinates. If instead
// we flip the image, then correcting left-right keypoint pairs requires a
// permutation on all the keypoints.
const flipPoseHorizontal = true;
canvas.width = videoWidth;
canvas.height = videoHeight;
async function poseDetectionFrame() {
if (guiState.changeToArchitecture) {
// Important to purge variables and free up GPU memory
guiState.net.dispose();
toggleLoadingUI(true);
guiState.net = await posenet.load({
architecture: guiState.changeToArchitecture,
outputStride: guiState.outputStride,
inputResolution: guiState.inputResolution,
multiplier: guiState.multiplier,
});
toggleLoadingUI(false);
guiState.architecture = guiState.changeToArchitecture;
guiState.changeToArchitecture = null;
}
if (guiState.changeToMultiplier) {
guiState.net.dispose();
toggleLoadingUI(true);
guiState.net = await posenet.load({
architecture: guiState.architecture,
outputStride: guiState.outputStride,
inputResolution: guiState.inputResolution,
multiplier: +guiState.changeToMultiplier,
quantBytes: guiState.quantBytes
});
toggleLoadingUI(false);
guiState.multiplier = +guiState.changeToMultiplier;
guiState.changeToMultiplier = null;
}
if (guiState.changeToOutputStride) {
// Important to purge variables and free up GPU memory
guiState.net.dispose();
toggleLoadingUI(true);
guiState.net = await posenet.load({
architecture: guiState.architecture,
outputStride: +guiState.changeToOutputStride,
inputResolution: guiState.inputResolution,
multiplier: guiState.multiplier,
quantBytes: guiState.quantBytes
});
toggleLoadingUI(false);
guiState.outputStride = +guiState.changeToOutputStride;
guiState.changeToOutputStride = null;
}
if (guiState.changeToInputResolution) {
// Important to purge variables and free up GPU memory
guiState.net.dispose();
toggleLoadingUI(true);
guiState.net = await posenet.load({
architecture: guiState.architecture,
outputStride: guiState.outputStride,
inputResolution: +guiState.changeToInputResolution,
multiplier: guiState.multiplier,
quantBytes: guiState.quantBytes
});
toggleLoadingUI(false);
guiState.inputResolution = +guiState.changeToInputResolution;
guiState.changeToInputResolution = null;
}
if (guiState.changeToQuantBytes) {
// Important to purge variables and free up GPU memory
guiState.net.dispose();
toggleLoadingUI(true);
guiState.net = await posenet.load({
architecture: guiState.architecture,
outputStride: guiState.outputStride,
inputResolution: guiState.inputResolution,
multiplier: guiState.multiplier,
quantBytes: guiState.changeToQuantBytes
});
toggleLoadingUI(false);
guiState.quantBytes = guiState.changeToQuantBytes;
guiState.changeToQuantBytes = null;
}
// Begin monitoring code for frames per second
stats.begin();
let poses = [];
let minPoseConfidence;
let minPartConfidence;
switch (guiState.algorithm) {
case 'single-pose':
const pose = await guiState.net.estimatePoses(video, {
flipHorizontal: flipPoseHorizontal,
decodingMethod: 'single-person'
});
poses = poses.concat(pose);
minPoseConfidence = +guiState.singlePoseDetection.minPoseConfidence;
minPartConfidence = +guiState.singlePoseDetection.minPartConfidence;
break;
case 'multi-pose':
let all_poses = await guiState.net.estimatePoses(video, {
flipHorizontal: flipPoseHorizontal,
decodingMethod: 'multi-person',
maxDetections: guiState.multiPoseDetection.maxPoseDetections,
scoreThreshold: guiState.multiPoseDetection.minPartConfidence,
nmsRadius: guiState.multiPoseDetection.nmsRadius
});
poses = poses.concat(all_poses);
minPoseConfidence = +guiState.multiPoseDetection.minPoseConfidence;
minPartConfidence = +guiState.multiPoseDetection.minPartConfidence;
break;
}
ctx.clearRect(0, 0, videoWidth, videoHeight);
if (guiState.output.showVideo) {
ctx.save();
ctx.scale(-1, 1);
ctx.translate(-videoWidth, 0);
ctx.drawImage(video, 0, 0, videoWidth, videoHeight);
ctx.restore();
}
// For each pose (i.e. person) detected in an image, loop through the poses
// and draw the resulting skeleton and keypoints if over certain confidence
// scores
poses.forEach(({score, keypoints}) => {
if (score >= minPoseConfidence) {
if (guiState.output.showPoints) {
//console.log(guiState.activity.Repetitions);
console.log('Repetitions: ', rep_count);
rep_count = rep_count + drawKeypoints(keypoints, minPartConfidence, ctx,
rep_count, wko_started, guiState.activity.Activity);
if (wko_started == 1 && done == 0){
ctx.font = "25px Arial";
ctx.fillText('Repetitions: ' + Math.round(rep_count/2), 10, 90,);
ctx.fillText(guiState.activity.Activity + ' workout has started', 100, 490,);
if (timer == 0){
if (count_down == 0){
t0 = new Date() / 1000;
count_down = 1;
}
ctx.font = "100px Arial";
var now = new Date() / 1000;
if (now - t0 <= 5){
ctx.fillText(5 - Math.round(now - t0), 250, 250,);
now = new Date() / 1000;
}
else{
timer = 1;
t0 = new Date() / 1000;
rep_count = 0;
}
}
else{
act_dur = Math.round(new Date() / 1000 - t0)
// console.log(new Date() / 1000)
console.log("Actual duration: ", act_dur);
console.log("guiState duration: ", guiState.activity.Duration_min);
console.log("Equal? ", act_dur == guiState.activity.Duration_min);
ctx.fillText('Duration: ' + act_dur + 's', 430, 90,);
if (guiState.activity.Activity == 'Weight Lifting' && Math.round(rep_count/2) == guiState.activity.Repetitions){
done = 1
}
else if (guiState.activity.Activity == 'Jumping jack' && act_dur == guiState.activity.Duration_min){
done = 1
}
}
}
else if(done == 1) {
if (wko_started == 1){
last_rep = rep_count;
}
wko_started = 0;
ctx.font = "45px Arial";
ctx.fillStyle = "green";
ctx.fillText('DONE!!', 210, 250,);
if (guiState.activity.Activity == 'Weight Lifting'){
ctx.fillText('Duration: ' + act_dur + 's', 160, 300,);
rep_count = 0;
global_zero();
}
else if (guiState.activity.Activity == 'Jumping jack'){
ctx.fillText('Repetitions: ' + Math.round(last_rep/2), 160, 300,);
rep_count = 0;
global_zero();
}
timer = 0;
}
else if (rep_count != guiState.activity.Repetitions || done == 0){
wko_started = 0;
rep_count = 0;
global_zero();
ctx.font = "29px Arial";
ctx.fillText('Workout not started, plase start a new train...', 10, 30,);
}
}
if (guiState.output.showSkeleton) {
drawSkeleton(keypoints, minPartConfidence, ctx);
}
if (guiState.output.showBoundingBox) {
drawBoundingBox(keypoints, ctx);
}
}
});
// End monitoring code for frames per second
stats.end();
requestAnimationFrame(poseDetectionFrame);
}
poseDetectionFrame();
}
/**
* Kicks off the demo by loading the posenet model, finding and loading
* available camera devices, and setting off the detectPoseInRealTime function.
*/
export async function bindPage() {
toggleLoadingUI(true);
const net = await posenet.load({
architecture: guiState.input.architecture,
outputStride: guiState.input.outputStride,
inputResolution: guiState.input.inputResolution,
multiplier: guiState.input.multiplier,
quantBytes: guiState.input.quantBytes
});
toggleLoadingUI(false);
let video;
try {
video = await loadVideo();
} catch (e) {
let info = document.getElementById('info');
info.textContent = 'this browser does not support video capture,' +
'or this device does not have a camera';
info.style.display = 'block';
throw e;
}
setupGui([], net);
//setupFPS();
detectPoseInRealTime(video, net);
}
navigator.getUserMedia = navigator.getUserMedia ||
navigator.webkitGetUserMedia || navigator.mozGetUserMedia;
// kick off the demo
bindPage();