Update index.js

index.js

@@ -1,4 +1,4 @@
-import { pipeline, env } from 'https://cdn.jsdelivr.net/npm/@xenova/transformers@2.10.1';
+import { MgpstrForSceneTextRecognition, MgpstrProcessor, RawImage } from 'https://cdn.jsdelivr.net/npm/@xenova/transformers@3.1.0';
 
 // Since we will download the model from the Hugging Face Hub, we can skip the local model check
 env.allowLocalModels = false;
@@ -9,13 +9,16 @@ const fileUpload = document.getElementById('upload');
 const imageContainer = document.getElementById('container');
 const example = document.getElementById('example');
 
-const EXAMPLE_URL = 'https://huggingface.co/datasets/Xenova/transformers.js-docs/resolve/main/city-streets.jpg';
-
-// Create a new object detection pipeline
+// Load Model
 status.textContent = 'Loading model...';
-const detector = await pipeline('object-detection', 'Xenova/detr-resnet-50');
+const model_id = 'onnx-community/mgp-str-base';
+const model = await MgpstrForSceneTextRecognition.from_pretrained(model_id);
+const processor = await MgpstrProcessor.from_pretrained(model_id);
 status.textContent = 'Ready';
 
+// Load image from the IIIT-5k dataset
+const EXAMPLE_URL = "https://i.postimg.cc/ZKwLg2Gw/367-14.png";
+
 example.addEventListener('click', (e) => {
     e.preventDefault();
     detect(EXAMPLE_URL);
@@ -42,38 +45,15 @@ async function detect(img) {
     imageContainer.style.backgroundImage = `url(${img})`;
 
     status.textContent = 'Analysing...';
-    const output = await detector(img, {
-        threshold: 0.5,
-        percentage: true,
-    });
-    status.textContent = '';
-    output.forEach(renderBox);
-}
-
-// Render a bounding box and label on the image
-function renderBox({ box, label }) {
-    const { xmax, xmin, ymax, ymin } = box;
-
-    // Generate a random color for the box
-    const color = '#' + Math.floor(Math.random() * 0xFFFFFF).toString(16).padStart(6, 0);
+    const image = await RawImage.read(img)-;
 
-    // Draw the box
-    const boxElement = document.createElement('div');
-    boxElement.className = 'bounding-box';
-    Object.assign(boxElement.style, {
-        borderColor: color,
-        left: 100 * xmin + '%',
-        top: 100 * ymin + '%',
-        width: 100 * (xmax - xmin) + '%',
-        height: 100 * (ymax - ymin) + '%',
-    })
+    // Preprocess the image
+    const result = await processor(image);
 
-    // Draw label
-    const labelElement = document.createElement('span');
-    labelElement.textContent = label;
-    labelElement.className = 'bounding-box-label';
-    labelElement.style.backgroundColor = color;
+    // Perform inference
+    const outputs = await model(result);
 
-    boxElement.appendChild(labelElement);
-    imageContainer.appendChild(boxElement);
-}
+    // Decode the model outputs
+    const generated_text = processor.batch_decode(outputs.logits).generated_text;
+    status.textContent = generated_text;
+}