implemented drag & drop for image processing

.gitignore

@@ -1,3 +1,3 @@
 node_modules
-private
 pnpm-lock.yaml
+samples

.npmignore

@@ -1,3 +1,6 @@
 node_modules
-private
 pnpm-lock.yaml
+samples
+typedoc
+test
+wiki

README.md

@@ -23,7 +23,13 @@ JavaScript module using TensorFlow/JS Machine Learning library
   Compatible with both software *tfjs-node* and  
   GPU accelerated backends *tfjs-node-gpu* using CUDA libraries  
 
-Check out [**Live Demo**](https://vladmandic.github.io/human/demo/index.html) for processing of live WebCam video or static images  
+<br>
+
+Check out [**Live Demo**](https://vladmandic.github.io/human/demo/index.html) app for processing of live WebCam video or static images  
+
+- To start video detection, simply press *Play*  
+- To process images, simply drag & drop in your Browser window  
+- Note: For optimal performance, select only models you'd like to use
 
 <br>
 
@@ -89,20 +95,30 @@ All options as presented in the demo application...
 
 <br>
 
-**Validation image:**  
+**Face Close-up:**  
-> [demo/index.html](demo/index.html?image=%22../assets/human-sample-upper.jpg%22)
+![Face](assets/screenshot-face.jpg)
 
-![Example Training Image](assets/screenshot-sample.png)
+<br>
 
-**Using static images:**  
+**Face under a high angle:**  
-> [demo/index.html](demo/index.html?images=true)
+![Angle](assets/screenshot-angle.jpg)
 
-![Example Using Image](assets/screenshot-images.jpg)
+<br>
 
-**Live WebCam view:**  
+**Full Person Details:**  
-> [demo/index.html](demo/index.html)
+![Pose](assets/screenshot-person.jpg)
 
-![Example Using WebCam](assets/screenshot-webcam.jpg)
+<br>
+
+**Pose Detection:**  
+![Pose](assets/screenshot-pose.jpg)
+
+<br>
+
+**Large Group:**  
+![Group](assets/screenshot-group.jpg)
+
+<br>
 
 **Face Similarity Matching:**  
 Extracts all faces from provided input images,  
@@ -112,11 +128,15 @@ and optionally matches detected face with database of known people to guess thei
 
 ![Face Matching](assets/screenshot-facematch.jpg)
 
+<br>
+
 **Face3D OpenGL Rendering:**  
 > [demo/face3d.html](demo/face3d.html)
 
 ![Face Matching](assets/screenshot-face3d.jpg)
 
+<br>
+
 **468-Point Face Mesh Defails:**  
 (view in full resolution to see keypoints)  
 

demo/index.js

@@ -77,6 +77,7 @@ const ui = {
   modelsPreload: true, // preload human models on startup
   modelsWarmup: true, // warmup human models on startup
   buffered: true, // should output be buffered between frames
+  interpolated: true, // should output be interpolated for smoothness between frames
   iconSize: '48px', // ui icon sizes
 
   // internal variables
@@ -228,8 +229,12 @@ async function drawResults(input) {
   }
 
   // draw all results using interpolated results
-  const interpolated = human.next(result);
+  if (ui.interpolated) {
-  human.draw.all(canvas, interpolated, drawOptions);
+    const interpolated = human.next(result);
+    human.draw.all(canvas, interpolated, drawOptions);
+  } else {
+    human.draw.all(canvas, result, drawOptions);
+  }
   /* alternatively use individual functions
   human.draw.face(canvas, result.face);
   human.draw.body(canvas, result.body);
@@ -246,20 +251,21 @@ async function drawResults(input) {
   const gpu = engine.backendInstance ? `gpu: ${(engine.backendInstance.numBytesInGPU ? engine.backendInstance.numBytesInGPU : 0).toLocaleString()} bytes` : '';
   const memory = `system: ${engine.state.numBytes.toLocaleString()} bytes ${gpu} | tensors: ${engine.state.numTensors.toLocaleString()}`;
   const processing = result.canvas ? `processing: ${result.canvas.width} x ${result.canvas.height}` : '';
-  const avgDetect = Math.trunc(10 * ui.detectFPS.reduce((a, b) => a + b, 0) / ui.detectFPS.length) / 10;
+  const avgDetect = ui.detectFPS.length > 0 ? Math.trunc(10 * ui.detectFPS.reduce((a, b) => a + b, 0) / ui.detectFPS.length) / 10 : 0;
-  const avgDraw = Math.trunc(10 * ui.drawFPS.reduce((a, b) => a + b, 0) / ui.drawFPS.length) / 10;
+  const avgDraw = ui.drawFPS.length > 0 ? Math.trunc(10 * ui.drawFPS.reduce((a, b) => a + b, 0) / ui.drawFPS.length) / 10 : 0;
   const warning = (ui.detectFPS.length > 5) && (avgDetect < 5) ? '<font color="lightcoral">warning: your performance is low: try switching to higher performance backend, lowering resolution or disabling some models</font>' : '';
+  const fps = avgDetect > 0 ? `FPS process:${avgDetect} refresh:${avgDraw}` : '';
   document.getElementById('log').innerHTML = `
     video: ${ui.camera.name} | facing: ${ui.camera.facing} | screen: ${window.innerWidth} x ${window.innerHeight} camera: ${ui.camera.width} x ${ui.camera.height} ${processing}<br>
     backend: ${human.tf.getBackend()} | ${memory}<br>
-    performance: ${str(lastDetectedResult.performance)}ms FPS process:${avgDetect} refresh:${avgDraw}<br>
+    performance: ${str(lastDetectedResult.performance)}ms ${fps}<br>
     ${warning}<br>
   `;
   ui.framesDraw++;
   ui.lastFrame = performance.now();
   // if buffered, immediate loop but limit frame rate although it's going to run slower as JS is singlethreaded
   if (ui.buffered) {
-    ui.drawThread = requestAnimationFrame(() => drawResults(input, canvas));
+    ui.drawThread = requestAnimationFrame(() => drawResults(input));
   } else {
     log('stopping buffered refresh');
     if (ui.drawThread) cancelAnimationFrame(ui.drawThread);
@@ -431,7 +437,7 @@ function runHumanDetect(input, canvas, timestamp) {
     ctx.drawImage(input, 0, 0, canvas.width, canvas.height);
     const data = ctx.getImageData(0, 0, canvas.width, canvas.height);
     // perform detection in worker
-    webWorker(input, data, canvas, userConfig, timestamp);
+    webWorker(input, data, canvas, timestamp);
     status();
   } else {
     human.detect(input, userConfig).then((result) => {
@@ -457,32 +463,66 @@ function runHumanDetect(input, canvas, timestamp) {
 }
 
 // main processing function when input is image, can use direct invocation or web worker
-async function processImage(input) {
+async function processImage(input, title) {
   return new Promise((resolve) => {
     const image = new Image();
+    image.onerror = async () => status('image loading error');
     image.onload = async () => {
-      log('processing image:', encodeURI(image.src));
+      ui.interpolated = false; // stop interpolating results if input is image
+      status(`processing image: ${title}`);
       const canvas = document.getElementById('canvas');
       image.width = image.naturalWidth;
       image.height = image.naturalHeight;
       canvas.width = human.config.filter.width && human.config.filter.width > 0 ? human.config.filter.width : image.naturalWidth;
       canvas.height = human.config.filter.height && human.config.filter.height > 0 ? human.config.filter.height : image.naturalHeight;
+      const origCacheSensitiry = userConfig.cacheSensitivity;
+      userConfig.cacheSensitivity = 0;
       const result = await human.detect(image, userConfig);
+      userConfig.cacheSensitivity = origCacheSensitiry;
       lastDetectedResult = result;
       await drawResults(image);
       const thumb = document.createElement('canvas');
       thumb.className = 'thumbnail';
-      thumb.width = window.innerWidth / (ui.columns + 0.1);
+      thumb.width = ui.columns > 1 ? window.innerWidth / (ui.columns + 0.1) : window.innerWidth - 14;
       thumb.height = thumb.width * canvas.height / canvas.width;
       if (result.face && result.face.length > 0) {
         thumb.title = result.face.map((a, i) => `#${i} face: ${Math.trunc(100 * a.faceScore)}% box: ${Math.trunc(100 * a.boxScore)}% age: ${Math.trunc(a.age)} gender: ${Math.trunc(100 * a.genderScore)}% ${a.gender}`).join(' | ');
       } else {
         thumb.title = 'no face detected';
       }
+      thumb.addEventListener('click', (evt) => {
+        const stdWidth = ui.columns > 1 ? window.innerWidth / (ui.columns + 0.1) : window.innerWidth - 14;
+        // zoom in/out on click
+        if (evt.target.style.width === `${stdWidth}px`) {
+          evt.target.style.width = '';
+          evt.target.style.height = `${document.getElementById('log').offsetTop - document.getElementById('media').offsetTop}px`;
+        } else {
+          evt.target.style.width = `${stdWidth}px`;
+          evt.target.style.height = '';
+        }
+        // copy to clipboard on click
+        if (typeof ClipboardItem !== 'undefined' && navigator.clipboard) {
+          evt.target.toBlob((blob) => {
+            // eslint-disable-next-line no-undef
+            const item = new ClipboardItem({ 'image/png': blob });
+            navigator.clipboard.write([item]);
+            log('copied image to clipboard');
+          });
+        }
+      });
       const ctx = thumb.getContext('2d');
       ctx.drawImage(canvas, 0, 0, canvas.width, canvas.height, 0, 0, thumb.width, thumb.height);
-      document.getElementById('samples-container').appendChild(thumb);
+      const prev = document.getElementsByClassName('thumbnail');
-      image.src = '';
+      if (prev && prev.length > 0) document.getElementById('samples-container').insertBefore(thumb, prev[0]);
+      else document.getElementById('samples-container').appendChild(thumb);
+
+      // finish up
+      status();
+      document.getElementById('play').style.display = 'none';
+      document.getElementById('loader').style.display = 'none';
+      if (ui.detectThread) cancelAnimationFrame(ui.detectThread);
+      if (ui.drawThread) cancelAnimationFrame(ui.drawThread);
+
       resolve(true);
     };
     image.src = input;
@@ -522,11 +562,7 @@ async function detectSampleImages() {
   status('processing images');
   document.getElementById('samples-container').innerHTML = '';
   for (const m of Object.values(menu)) m.hide();
-  for (const image of ui.samples) await processImage(image);
+  for (const image of ui.samples) await processImage(image, image);
-  status();
-  document.getElementById('play').style.display = 'none';
-  document.getElementById('loader').style.display = 'none';
-  if (ui.detectThread) cancelAnimationFrame(ui.detectThread);
 }
 
 function setupMenu() {
@@ -604,8 +640,8 @@ function setupMenu() {
     human.config.hand.rotation = val;
   });
   menu.process.addHTML('<hr style="border-style: inset; border-color: dimgray">');
-  menu.process.addButton('process sample images', 'process images', () => detectSampleImages());
+  // menu.process.addButton('process sample images', 'process images', () => detectSampleImages());
-  menu.process.addHTML('<hr style="border-style: inset; border-color: dimgray">');
+  // menu.process.addHTML('<hr style="border-style: inset; border-color: dimgray">');
   menu.process.addChart('FPS', 'FPS');
 
   menu.models = new Menu(document.body, '', { top, left: x[3] });
@@ -676,6 +712,31 @@ async function drawWarmup(res) {
   await human.draw.all(canvas, res, drawOptions);
 }
 
+async function processDataURL(f) {
+  return new Promise((resolve) => {
+    const reader = new FileReader();
+    reader.onload = async (e) => {
+      const dataURL = e.target.result;
+      await processImage(dataURL, f.name);
+      document.getElementById('canvas').style.display = 'none';
+      resolve(true);
+    };
+    reader.readAsDataURL(f);
+  });
+}
+
+async function dragAndDrop() {
+  document.body.addEventListener('dragenter', (evt) => evt.preventDefault());
+  document.body.addEventListener('dragleave', (evt) => evt.preventDefault());
+  document.body.addEventListener('dragover', (evt) => evt.preventDefault());
+  document.body.addEventListener('drop', async (evt) => {
+    evt.preventDefault();
+    evt.dataTransfer.dropEffect = 'copy';
+    if (evt.dataTransfer.files.length < 2) ui.columns = 1;
+    for (const f of evt.dataTransfer.files) await processDataURL(f);
+  });
+}
+
 async function pwaRegister() {
   if (!pwa.enabled) return;
   if ('serviceWorker' in navigator) {
@@ -790,11 +851,16 @@ async function main() {
   document.getElementById('play').style.display = 'block';
   for (const m of Object.values(menu)) m.hide();
 
+  // init drag & drop
+
+  await dragAndDrop();
+
   if (params.has('image')) {
     try {
       const image = JSON.parse(params.get('image'));
       log('overriding image:', image);
       ui.samples = [image];
+      ui.columns = 1;
     } catch {
       status('cannot parse input image');
       log('cannot parse input image', params.get('image'));

dist/human.node-gpu.js

@@ -155,8 +155,8 @@ var config = {
     enabled: true,
     detector: {
       modelPath: "blazeface.json",
-      rotation: false,
+      rotation: true,
-      maxDetected: 10,
+      maxDetected: 5,
       skipFrames: 15,
       minConfidence: 0.2,
       iouThreshold: 0.1,
@@ -9806,7 +9806,7 @@ __export(draw_exports, {
   person: () => person
 });
 var options = {
-  color: "rgba(173, 216, 230, 0.3)",
+  color: "rgba(173, 216, 230, 0.6)",
   labelColor: "rgba(173, 216, 230, 1)",
   shadowColor: "black",
   font: 'small-caps 14px "Segoe UI"',

dist/human.node-wasm.js

@@ -156,8 +156,8 @@ var config = {
     enabled: true,
     detector: {
       modelPath: "blazeface.json",
-      rotation: false,
+      rotation: true,
-      maxDetected: 10,
+      maxDetected: 5,
       skipFrames: 15,
       minConfidence: 0.2,
       iouThreshold: 0.1,
@@ -9807,7 +9807,7 @@ __export(draw_exports, {
   person: () => person
 });
 var options = {
-  color: "rgba(173, 216, 230, 0.3)",
+  color: "rgba(173, 216, 230, 0.6)",
   labelColor: "rgba(173, 216, 230, 1)",
   shadowColor: "black",
   font: 'small-caps 14px "Segoe UI"',

dist/human.node.js

@@ -155,8 +155,8 @@ var config = {
     enabled: true,
     detector: {
       modelPath: "blazeface.json",
-      rotation: false,
+      rotation: true,
-      maxDetected: 10,
+      maxDetected: 5,
       skipFrames: 15,
       minConfidence: 0.2,
       iouThreshold: 0.1,
@@ -9806,7 +9806,7 @@ __export(draw_exports, {
   person: () => person
 });
 var options = {
-  color: "rgba(173, 216, 230, 0.3)",
+  color: "rgba(173, 216, 230, 0.6)",
   labelColor: "rgba(173, 216, 230, 1)",
   shadowColor: "black",
   font: 'small-caps 14px "Segoe UI"',

src/config.ts

@@ -241,10 +241,10 @@ const config: Config = {
                              // (note: module is not loaded until it is required)
     detector: {
       modelPath: 'blazeface.json', // detector model, can be absolute path or relative to modelBasePath
-      rotation: false,       // use best-guess rotated face image or just box with rotation as-is
+      rotation: true,        // use best-guess rotated face image or just box with rotation as-is
                              // false means higher performance, but incorrect mesh mapping if face angle is above 20 degrees
                              // this parameter is not valid in nodejs
-      maxDetected: 10,          // maximum number of faces detected in the input
+      maxDetected: 5,        // maximum number of faces detected in the input
                              // should be set to the minimum number for performance
       skipFrames: 15,        // how many max frames to go without re-running the face bounding box detector
                              // only used when cacheSensitivity is not zero

src/draw/draw.ts

@@ -47,7 +47,7 @@ export interface DrawOptions {
 }
 
 export const options: DrawOptions = {
-  color: <string>'rgba(173, 216, 230, 0.3)', // 'lightblue' with light alpha channel
+  color: <string>'rgba(173, 216, 230, 0.6)', // 'lightblue' with light alpha channel
   labelColor: <string>'rgba(173, 216, 230, 1)', // 'lightblue' with dark alpha channel
   shadowColor: <string>'black',
   font: <string>'small-caps 14px "Segoe UI"',