New Case Study:See how Anthropic automated 95% of dependency reviews with Socket.Learn More →
Socket
Sign inDemoInstall
Socket
Sign inDemoInstall

cycle-posenet-driver

Package Overview
Dependencies
Maintainers
1
Versions
35
Alerts
File Explorer

Advanced tools

License
Socket logo

Install Socket

Detect and block malicious and high-risk dependencies

Install

cycle-posenet-driver - npm Package Compare versions

Comparing version 0.0.1 to 0.0.2

4

lib/cjs/pose_detection.d.ts

@@ -5,5 +5,5 @@ export declare function makePoseDetectionDriver({videoWidth, videoHeight, flipHorizontal}?: {

flipHorizontal?: boolean;
}): (sink$: any) => {
}): (sink$: any) => Promise<{
DOM: any;
poses: any;
};
}>;

287

lib/cjs/pose_detection.js

@@ -138,103 +138,102 @@ "use strict";

return function poseDetectionDriver(sink$) {
var _this = this;
var params = null;
var video = null;
var context = null;
var posesListener = null;
var poses$ = xstream_1.default.create({
start: function (listener) {
posesListener = function (result) {
listener.next(result);
};
},
stop: function () {
posesListener = null;
},
});
function poseDetectionFrame() {
return __awaiter(this, void 0, void 0, function () {
var _a, imageScaleFactor, outputStride, poses, minPoseConfidence, minPartConfidence, _b, pose, outPoses;
return __generator(this, function (_c) {
switch (_c.label) {
case 0:
if (!params.changeToArchitecture) return [3 /*break*/, 2];
// Important to purge variables and free up GPU memory
params.net.dispose();
// Load the PoseNet model weights for either the 0.50, 0.75, 1.00, or
// 1.01 version
_a = params;
return [4 /*yield*/, posenet.load(+params.changeToArchitecture)];
case 1:
// Load the PoseNet model weights for either the 0.50, 0.75, 1.00, or
// 1.01 version
_a.net = _c.sent();
params.changeToArchitecture = null;
_c.label = 2;
case 2:
// Begin monitoring code for frames per second
stats.begin();
imageScaleFactor = params.input.imageScaleFactor;
outputStride = +params.input.outputStride;
poses = [];
_b = params.algorithm;
switch (_b) {
case 'single-pose': return [3 /*break*/, 3];
case 'multi-pose': return [3 /*break*/, 5];
}
return [3 /*break*/, 7];
case 3: return [4 /*yield*/, params.net.estimateSinglePose(video, imageScaleFactor, flipHorizontal, outputStride)];
case 4:
pose = _c.sent();
poses.push(pose);
minPoseConfidence = +params.singlePoseDetection.minPoseConfidence;
minPartConfidence = +params.singlePoseDetection.minPartConfidence;
return [3 /*break*/, 7];
case 5: return [4 /*yield*/, params.net.estimateMultiplePoses(video, imageScaleFactor, flipHorizontal, outputStride, params.multiPoseDetection.maxPoseDetections, params.multiPoseDetection.minPartConfidence, params.multiPoseDetection.nmsRadius)];
case 6:
poses = _c.sent();
minPoseConfidence = +params.multiPoseDetection.minPoseConfidence;
minPartConfidence = +params.multiPoseDetection.minPartConfidence;
return [3 /*break*/, 7];
case 7:
context.clearRect(0, 0, videoWidth, videoHeight);
if (params.output.showVideo) {
context.save();
context.scale(-1, 1);
context.translate(-videoWidth, 0);
context.drawImage(video, 0, 0, videoWidth, videoHeight);
context.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(function (_a) {
var score = _a.score, keypoints = _a.keypoints;
if (score >= minPoseConfidence) {
if (params.output.showPoints) {
utils_1.drawKeypoints(keypoints, minPartConfidence, context);
return __awaiter(this, void 0, void 0, function () {
function poseDetectionFrame() {
return __awaiter(this, void 0, void 0, function () {
var _a, imageScaleFactor, outputStride, poses, minPoseConfidence, minPartConfidence, _b, pose, outPoses;
return __generator(this, function (_c) {
switch (_c.label) {
case 0:
if (!params.changeToArchitecture) return [3 /*break*/, 2];
// Important to purge variables and free up GPU memory
params.net.dispose();
// Load the PoseNet model weights for either the 0.50, 0.75, 1.00, or
// 1.01 version
_a = params;
return [4 /*yield*/, posenet.load(+params.changeToArchitecture)];
case 1:
// Load the PoseNet model weights for either the 0.50, 0.75, 1.00, or
// 1.01 version
_a.net = _c.sent();
params.changeToArchitecture = null;
_c.label = 2;
case 2:
// Begin monitoring code for frames per second
stats.begin();
imageScaleFactor = params.input.imageScaleFactor;
outputStride = +params.input.outputStride;
poses = [];
_b = params.algorithm;
switch (_b) {
case 'single-pose': return [3 /*break*/, 3];
case 'multi-pose': return [3 /*break*/, 5];
}
return [3 /*break*/, 7];
case 3: return [4 /*yield*/, params.net.estimateSinglePose(video, imageScaleFactor, flipHorizontal, outputStride)];
case 4:
pose = _c.sent();
poses.push(pose);
minPoseConfidence = +params.singlePoseDetection.minPoseConfidence;
minPartConfidence = +params.singlePoseDetection.minPartConfidence;
return [3 /*break*/, 7];
case 5: return [4 /*yield*/, params.net.estimateMultiplePoses(video, imageScaleFactor, flipHorizontal, outputStride, params.multiPoseDetection.maxPoseDetections, params.multiPoseDetection.minPartConfidence, params.multiPoseDetection.nmsRadius)];
case 6:
poses = _c.sent();
minPoseConfidence = +params.multiPoseDetection.minPoseConfidence;
minPartConfidence = +params.multiPoseDetection.minPartConfidence;
return [3 /*break*/, 7];
case 7:
context.clearRect(0, 0, videoWidth, videoHeight);
if (params.output.showVideo) {
context.save();
context.scale(-1, 1);
context.translate(-videoWidth, 0);
context.drawImage(video, 0, 0, videoWidth, videoHeight);
context.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(function (_a) {
var score = _a.score, keypoints = _a.keypoints;
if (score >= minPoseConfidence) {
if (params.output.showPoints) {
utils_1.drawKeypoints(keypoints, minPartConfidence, context);
}
if (params.output.showSkeleton) {
utils_1.drawSkeleton(keypoints, minPartConfidence, context);
}
}
if (params.output.showSkeleton) {
utils_1.drawSkeleton(keypoints, minPartConfidence, context);
}
});
outPoses = poses
.filter(function (pose) { return pose.score >= minPoseConfidence; })
.map(function (pose) { return (__assign({}, pose, { keypoints: pose.keypoints.filter(function (keypoint) { return keypoint.score >= minPartConfidence; }) })); });
if (posesListener) {
posesListener(outPoses);
// End monitoring code for frames per second
stats.end();
requestAnimationFrame(poseDetectionFrame);
}
});
outPoses = poses
.filter(function (pose) { return pose.score >= minPoseConfidence; })
.map(function (pose) { return (__assign({}, pose, { keypoints: pose.keypoints.filter(function (keypoint) { return keypoint.score >= minPartConfidence; }) })); });
if (posesListener) {
posesListener(outPoses);
// End monitoring code for frames per second
stats.end();
requestAnimationFrame(poseDetectionFrame);
}
return [2 /*return*/];
}
return [2 /*return*/];
}
});
});
});
}
window.addEventListener('load', function () { return __awaiter(_this, void 0, void 0, function () {
var canvas, _a, gui;
}
var params, video, context, posesListener, poses$, canvas, _a, gui, initialParams, vdom$;
return __generator(this, function (_b) {
switch (_b.label) {
case 0:
params = null;
video = null;
context = null;
posesListener = null;
poses$ = xstream_1.default.create({
start: function (listener) {
posesListener = function (result) {
listener.next(result);
};
},
stop: function () {
posesListener = null;
},
});
if (!!video) return [3 /*break*/, 3];

@@ -264,54 +263,54 @@ return [4 /*yield*/, utils_1.setupCamera(document.querySelector('#pose_detection_video'), videoWidth, videoHeight)];

_b.label = 3;
case 3: return [2 /*return*/];
case 3:
initialParams = {
algorithm: 'multi-pose',
input: {
mobileNetArchitecture: utils_1.isMobile() ? '0.50' : '0.75',
outputStride: 16,
imageScaleFactor: 0.5,
},
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,
},
net: null,
changeToArchitecture: null,
};
sink$.fold(function (curParams, newParams) {
Object.keys(newParams).map(function (key) {
if (typeof newParams[key] === 'object') {
Object.assign(curParams[key], newParams[key]);
}
else {
curParams[key] = newParams[key];
}
return curParams;
});
return curParams;
}, initialParams).addListener({
next: function (newParams) {
params = newParams;
}
});
vdom$ = xstream_1.default.of((snabbdom_pragma_1.default.createElement("div", { id: "pose_detection", style: { position: "relative" } },
snabbdom_pragma_1.default.createElement("video", { id: "pose_detection_video", style: { display: 'none' }, autoPlay: true }),
snabbdom_pragma_1.default.createElement("canvas", { id: "pose_detection_canvas" }))));
return [2 /*return*/, {
DOM: adapt_1.adapt(vdom$),
poses: adapt_1.adapt(poses$),
}];
}
});
}); });
var initialParams = {
algorithm: 'multi-pose',
input: {
mobileNetArchitecture: utils_1.isMobile() ? '0.50' : '0.75',
outputStride: 16,
imageScaleFactor: 0.5,
},
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,
},
net: null,
changeToArchitecture: null,
};
sink$.fold(function (curParams, newParams) {
Object.keys(newParams).map(function (key) {
if (typeof newParams[key] === 'object') {
Object.assign(curParams[key], newParams[key]);
}
else {
curParams[key] = newParams[key];
}
return curParams;
});
return curParams;
}, initialParams).addListener({
next: function (newParams) {
params = newParams;
}
});
var vdom$ = xstream_1.default.of((snabbdom_pragma_1.default.createElement("div", { id: "pose_detection", style: { position: "relative" } },
snabbdom_pragma_1.default.createElement("video", { id: "pose_detection_video", style: { display: 'none' }, autoPlay: true }),
snabbdom_pragma_1.default.createElement("canvas", { id: "pose_detection_canvas" }))));
return {
DOM: adapt_1.adapt(vdom$),
poses: adapt_1.adapt(poses$),
};
};

@@ -318,0 +317,0 @@ }

4

lib/es6/pose_detection.d.ts

@@ -5,5 +5,5 @@ export declare function makePoseDetectionDriver({videoWidth, videoHeight, flipHorizontal}?: {

flipHorizontal?: boolean;
}): (sink$: any) => {
}): (sink$: any) => Promise<{
DOM: any;
poses: any;
};
}>;

287

lib/es6/pose_detection.js

@@ -126,103 +126,102 @@ var __assign = (this && this.__assign) || Object.assign || function(t) {

return function poseDetectionDriver(sink$) {
var _this = this;
var params = null;
var video = null;
var context = null;
var posesListener = null;
var poses$ = xs.create({
start: function (listener) {
posesListener = function (result) {
listener.next(result);
};
},
stop: function () {
posesListener = null;
},
});
function poseDetectionFrame() {
return __awaiter(this, void 0, void 0, function () {
var _a, imageScaleFactor, outputStride, poses, minPoseConfidence, minPartConfidence, _b, pose, outPoses;
return __generator(this, function (_c) {
switch (_c.label) {
case 0:
if (!params.changeToArchitecture) return [3 /*break*/, 2];
// Important to purge variables and free up GPU memory
params.net.dispose();
// Load the PoseNet model weights for either the 0.50, 0.75, 1.00, or
// 1.01 version
_a = params;
return [4 /*yield*/, posenet.load(+params.changeToArchitecture)];
case 1:
// Load the PoseNet model weights for either the 0.50, 0.75, 1.00, or
// 1.01 version
_a.net = _c.sent();
params.changeToArchitecture = null;
_c.label = 2;
case 2:
// Begin monitoring code for frames per second
stats.begin();
imageScaleFactor = params.input.imageScaleFactor;
outputStride = +params.input.outputStride;
poses = [];
_b = params.algorithm;
switch (_b) {
case 'single-pose': return [3 /*break*/, 3];
case 'multi-pose': return [3 /*break*/, 5];
}
return [3 /*break*/, 7];
case 3: return [4 /*yield*/, params.net.estimateSinglePose(video, imageScaleFactor, flipHorizontal, outputStride)];
case 4:
pose = _c.sent();
poses.push(pose);
minPoseConfidence = +params.singlePoseDetection.minPoseConfidence;
minPartConfidence = +params.singlePoseDetection.minPartConfidence;
return [3 /*break*/, 7];
case 5: return [4 /*yield*/, params.net.estimateMultiplePoses(video, imageScaleFactor, flipHorizontal, outputStride, params.multiPoseDetection.maxPoseDetections, params.multiPoseDetection.minPartConfidence, params.multiPoseDetection.nmsRadius)];
case 6:
poses = _c.sent();
minPoseConfidence = +params.multiPoseDetection.minPoseConfidence;
minPartConfidence = +params.multiPoseDetection.minPartConfidence;
return [3 /*break*/, 7];
case 7:
context.clearRect(0, 0, videoWidth, videoHeight);
if (params.output.showVideo) {
context.save();
context.scale(-1, 1);
context.translate(-videoWidth, 0);
context.drawImage(video, 0, 0, videoWidth, videoHeight);
context.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(function (_a) {
var score = _a.score, keypoints = _a.keypoints;
if (score >= minPoseConfidence) {
if (params.output.showPoints) {
drawKeypoints(keypoints, minPartConfidence, context);
return __awaiter(this, void 0, void 0, function () {
function poseDetectionFrame() {
return __awaiter(this, void 0, void 0, function () {
var _a, imageScaleFactor, outputStride, poses, minPoseConfidence, minPartConfidence, _b, pose, outPoses;
return __generator(this, function (_c) {
switch (_c.label) {
case 0:
if (!params.changeToArchitecture) return [3 /*break*/, 2];
// Important to purge variables and free up GPU memory
params.net.dispose();
// Load the PoseNet model weights for either the 0.50, 0.75, 1.00, or
// 1.01 version
_a = params;
return [4 /*yield*/, posenet.load(+params.changeToArchitecture)];
case 1:
// Load the PoseNet model weights for either the 0.50, 0.75, 1.00, or
// 1.01 version
_a.net = _c.sent();
params.changeToArchitecture = null;
_c.label = 2;
case 2:
// Begin monitoring code for frames per second
stats.begin();
imageScaleFactor = params.input.imageScaleFactor;
outputStride = +params.input.outputStride;
poses = [];
_b = params.algorithm;
switch (_b) {
case 'single-pose': return [3 /*break*/, 3];
case 'multi-pose': return [3 /*break*/, 5];
}
return [3 /*break*/, 7];
case 3: return [4 /*yield*/, params.net.estimateSinglePose(video, imageScaleFactor, flipHorizontal, outputStride)];
case 4:
pose = _c.sent();
poses.push(pose);
minPoseConfidence = +params.singlePoseDetection.minPoseConfidence;
minPartConfidence = +params.singlePoseDetection.minPartConfidence;
return [3 /*break*/, 7];
case 5: return [4 /*yield*/, params.net.estimateMultiplePoses(video, imageScaleFactor, flipHorizontal, outputStride, params.multiPoseDetection.maxPoseDetections, params.multiPoseDetection.minPartConfidence, params.multiPoseDetection.nmsRadius)];
case 6:
poses = _c.sent();
minPoseConfidence = +params.multiPoseDetection.minPoseConfidence;
minPartConfidence = +params.multiPoseDetection.minPartConfidence;
return [3 /*break*/, 7];
case 7:
context.clearRect(0, 0, videoWidth, videoHeight);
if (params.output.showVideo) {
context.save();
context.scale(-1, 1);
context.translate(-videoWidth, 0);
context.drawImage(video, 0, 0, videoWidth, videoHeight);
context.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(function (_a) {
var score = _a.score, keypoints = _a.keypoints;
if (score >= minPoseConfidence) {
if (params.output.showPoints) {
drawKeypoints(keypoints, minPartConfidence, context);
}
if (params.output.showSkeleton) {
drawSkeleton(keypoints, minPartConfidence, context);
}
}
if (params.output.showSkeleton) {
drawSkeleton(keypoints, minPartConfidence, context);
}
});
outPoses = poses
.filter(function (pose) { return pose.score >= minPoseConfidence; })
.map(function (pose) { return (__assign({}, pose, { keypoints: pose.keypoints.filter(function (keypoint) { return keypoint.score >= minPartConfidence; }) })); });
if (posesListener) {
posesListener(outPoses);
// End monitoring code for frames per second
stats.end();
requestAnimationFrame(poseDetectionFrame);
}
});
outPoses = poses
.filter(function (pose) { return pose.score >= minPoseConfidence; })
.map(function (pose) { return (__assign({}, pose, { keypoints: pose.keypoints.filter(function (keypoint) { return keypoint.score >= minPartConfidence; }) })); });
if (posesListener) {
posesListener(outPoses);
// End monitoring code for frames per second
stats.end();
requestAnimationFrame(poseDetectionFrame);
}
return [2 /*return*/];
}
return [2 /*return*/];
}
});
});
});
}
window.addEventListener('load', function () { return __awaiter(_this, void 0, void 0, function () {
var canvas, _a, gui;
}
var params, video, context, posesListener, poses$, canvas, _a, gui, initialParams, vdom$;
return __generator(this, function (_b) {
switch (_b.label) {
case 0:
params = null;
video = null;
context = null;
posesListener = null;
poses$ = xs.create({
start: function (listener) {
posesListener = function (result) {
listener.next(result);
};
},
stop: function () {
posesListener = null;
},
});
if (!!video) return [3 /*break*/, 3];

@@ -252,56 +251,56 @@ return [4 /*yield*/, setupCamera(document.querySelector('#pose_detection_video'), videoWidth, videoHeight)];

_b.label = 3;
case 3: return [2 /*return*/];
case 3:
initialParams = {
algorithm: 'multi-pose',
input: {
mobileNetArchitecture: isMobile() ? '0.50' : '0.75',
outputStride: 16,
imageScaleFactor: 0.5,
},
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,
},
net: null,
changeToArchitecture: null,
};
sink$.fold(function (curParams, newParams) {
Object.keys(newParams).map(function (key) {
if (typeof newParams[key] === 'object') {
Object.assign(curParams[key], newParams[key]);
}
else {
curParams[key] = newParams[key];
}
return curParams;
});
return curParams;
}, initialParams).addListener({
next: function (newParams) {
params = newParams;
}
});
vdom$ = xs.of((Snabbdom.createElement("div", { id: "pose_detection", style: { position: "relative" } },
Snabbdom.createElement("video", { id: "pose_detection_video", style: { display: 'none' }, autoPlay: true }),
Snabbdom.createElement("canvas", { id: "pose_detection_canvas" }))));
return [2 /*return*/, {
DOM: adapt(vdom$),
poses: adapt(poses$),
}];
}
});
}); });
var initialParams = {
algorithm: 'multi-pose',
input: {
mobileNetArchitecture: isMobile() ? '0.50' : '0.75',
outputStride: 16,
imageScaleFactor: 0.5,
},
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,
},
net: null,
changeToArchitecture: null,
};
sink$.fold(function (curParams, newParams) {
Object.keys(newParams).map(function (key) {
if (typeof newParams[key] === 'object') {
Object.assign(curParams[key], newParams[key]);
}
else {
curParams[key] = newParams[key];
}
return curParams;
});
return curParams;
}, initialParams).addListener({
next: function (newParams) {
params = newParams;
}
});
var vdom$ = xs.of((Snabbdom.createElement("div", { id: "pose_detection", style: { position: "relative" } },
Snabbdom.createElement("video", { id: "pose_detection_video", style: { display: 'none' }, autoPlay: true }),
Snabbdom.createElement("canvas", { id: "pose_detection_canvas" }))));
return {
DOM: adapt(vdom$),
poses: adapt(poses$),
};
};
}
//# sourceMappingURL=pose_detection.js.map

2

package.json
{
"name": "cycle-posenet-driver",
"version": "0.0.1",
"version": "0.0.2",
"description": "",

@@ -5,0 +5,0 @@ "author": "Michael Jae-Yoon Chung",

lib/cjs/pose_detection.js.map

Sorry, the diff of this file is not supported yet

lib/es6/pose_detection.js.map

Sorry, the diff of this file is not supported yet

src/pose_detection.tsx

Sorry, the diff of this file is not supported yet

SocketSocket SOC 2 Logo

Product

  • Package Alerts
  • Integrations
  • Docs
  • Pricing
  • FAQ
  • Roadmap
  • Changelog

About

Packages

npm

Go

Maven

PyPI

Rubygems

Stay in touch

Get open source security insights delivered straight into your inbox.

  • Terms
  • Privacy
  • Security

Made with ⚡️ by Socket Inc