Update README.md

README.md

@@ -3,4 +3,192 @@ library_name: transformers.js
 tags:
 - pose-estimation
 license: agpl-3.0
----
+---
+
+YOLOv8x-pose-p6 with ONNX weights to be compatible with Transformers.js.
+
+## Usage (Transformers.js)
+
+If you haven't already, you can install the [Transformers.js](https://huggingface.co/docs/transformers.js) JavaScript library from [NPM](https://www.npmjs.com/package/@xenova/transformers) using:
+```bash
+npm i @xenova/transformers
+```
+
+**Example:** Perform pose-estimation w/ `Xenova/yolov8x-pose-p6`.
+
+```js
+import { AutoModel, AutoProcessor, RawImage } from '@xenova/transformers';
+
+// Load model and processor
+const model_id = 'Xenova/yolov8x-pose-p6';
+const model = await AutoModel.from_pretrained(model_id);
+const processor = await AutoProcessor.from_pretrained(model_id);
+
+// Read image and run processor
+const url = 'https://huggingface.co/datasets/Xenova/transformers.js-docs/resolve/main/football-match.jpg';
+const image = await RawImage.read(url);
+const { pixel_values } = await processor(image);
+
+// Set thresholds
+const threshold = 0.3; // Remove detections with low confidence
+const iouThreshold = 0.5; // Used to remove duplicates
+const pointThreshold = 0.3; // Hide uncertain points
+
+// Predict bounding boxes and keypoints
+const { output0 } = await model({ images: pixel_values });
+
+// Post-process:
+const permuted = output0[0].transpose(1, 0);
+// `permuted` is a Tensor of shape [ 8400, 56 ]:
+// - 8400 potential detections
+// - 56 parameters for each box:
+//   - 4 for the bounding box dimensions (x-center, y-center, width, height)
+//   - 1 for the confidence score
+//   - 17 * 3 = 51 for the pose keypoints: 17 labels, each with (x, y, visibilitiy)
+
+// Example code to format it nicely:
+const results = [];
+const [scaledHeight, scaledWidth] = pixel_values.dims.slice(-2);
+for (const [xc, yc, w, h, score, ...keypoints] of permuted.tolist()) {
+    if (score < threshold) continue;
+
+    // Get pixel values, taking into account the original image size
+    const x1 = (xc - w / 2) / scaledWidth * image.width;
+    const y1 = (yc - h / 2) / scaledHeight * image.height;
+    const x2 = (xc + w / 2) / scaledWidth * image.width;
+    const y2 = (yc + h / 2) / scaledHeight * image.height;
+    results.push({ x1, x2, y1, y2, score, keypoints })
+}
+
+
+// Define helper functions
+function removeDuplicates(detections, iouThreshold) {
+    const filteredDetections = [];
+
+    for (const detection of detections) {
+        let isDuplicate = false;
+        let duplicateIndex = -1;
+        let maxIoU = 0;
+
+        for (let i = 0; i < filteredDetections.length; ++i) {
+            const filteredDetection = filteredDetections[i];
+            const iou = calculateIoU(detection, filteredDetection);
+            if (iou > iouThreshold) {
+                isDuplicate = true;
+                if (iou > maxIoU) {
+                    maxIoU = iou;
+                    duplicateIndex = i;
+                }
+            }
+        }
+
+        if (!isDuplicate) {
+            filteredDetections.push(detection);
+        } else if (duplicateIndex !== -1 && detection.score > filteredDetections[duplicateIndex].score) {
+            filteredDetections[duplicateIndex] = detection;
+        }
+    }
+
+    return filteredDetections;
+}
+
+function calculateIoU(detection1, detection2) {
+    const xOverlap = Math.max(0, Math.min(detection1.x2, detection2.x2) - Math.max(detection1.x1, detection2.x1));
+    const yOverlap = Math.max(0, Math.min(detection1.y2, detection2.y2) - Math.max(detection1.y1, detection2.y1));
+    const overlapArea = xOverlap * yOverlap;
+
+    const area1 = (detection1.x2 - detection1.x1) * (detection1.y2 - detection1.y1);
+    const area2 = (detection2.x2 - detection2.x1) * (detection2.y2 - detection2.y1);
+    const unionArea = area1 + area2 - overlapArea;
+
+    return overlapArea / unionArea;
+}
+
+const filteredResults = removeDuplicates(results, iouThreshold);
+
+// Display results
+for (const { x1, x2, y1, y2, score, keypoints } of filteredResults) {
+    console.log(`Found person at [${x1}, ${y1}, ${x2}, ${y2}] with score ${score.toFixed(3)}`)
+    for (let i = 0; i < keypoints.length; i += 3) {
+        const label = model.config.id2label[Math.floor(i / 3)];
+        const [x, y, point_score] = keypoints.slice(i, i + 3);
+        if (point_score < pointThreshold) continue;
+        console.log(`  - ${label}: (${x.toFixed(2)}, ${y.toFixed(2)}) with score ${point_score.toFixed(3)}`);
+    }
+}
+```
+
+<details>
+
+<summary>See example output</summary>
+
+```
+Found person at [535.95703125, 43.12074284553528, 644.3259429931641, 337.3436294078827] with score 0.760
+  - nose: (885.58, 179.72) with score 0.975
+  - left_eye: (897.09, 165.24) with score 0.976
+  - right_eye: (874.85, 164.54) with score 0.851
+  - left_ear: (914.39, 169.48) with score 0.806
+  - left_shoulder: (947.49, 252.34) with score 0.996
+  - right_shoulder: (840.67, 244.42) with score 0.665
+  - left_elbow: (1001.36, 351.66) with score 0.983
+  - left_wrist: (1011.84, 472.31) with score 0.954
+  - left_hip: (931.52, 446.28) with score 0.986
+  - right_hip: (860.66, 442.87) with score 0.828
+  - left_knee: (930.67, 625.64) with score 0.979
+  - right_knee: (872.17, 620.36) with score 0.735
+  - left_ankle: (929.01, 772.34) with score 0.880
+  - right_ankle: (882.23, 778.68) with score 0.454
+Found person at [0.4024791717529297, 59.50179467201233, 156.87244415283203, 370.64377751350406] with score 0.853
+  - nose: (115.39, 198.06) with score 0.918
+  - left_eye: (120.26, 177.71) with score 0.830
+  - right_eye: (105.47, 179.69) with score 0.757
+  - left_ear: (144.87, 185.18) with score 0.711
+  - right_ear: (97.69, 188.45) with score 0.468
+  - left_shoulder: (178.03, 268.88) with score 0.975
+  - right_shoulder: (80.69, 273.99) with score 0.954
+  - left_elbow: (203.06, 383.33) with score 0.923
+  - right_elbow: (43.32, 376.35) with score 0.856
+  - left_wrist: (215.74, 504.02) with score 0.888
+  - right_wrist: (6.77, 462.65) with score 0.812
+  - left_hip: (165.70, 473.24) with score 0.990
+  - right_hip: (97.84, 471.69) with score 0.986
+  - left_knee: (183.26, 646.61) with score 0.991
+  - right_knee: (104.04, 651.17) with score 0.989
+  - left_ankle: (199.88, 823.24) with score 0.966
+  - right_ankle: (104.66, 827.66) with score 0.963
+Found person at [107.49130249023438, 12.557352638244629, 501.3542175292969, 527.4827188491821] with score 0.872
+  - nose: (246.06, 180.81) with score 0.722
+  - left_eye: (236.99, 148.85) with score 0.523
+  - left_ear: (289.26, 152.23) with score 0.770
+  - left_shoulder: (391.63, 256.55) with score 0.992
+  - right_shoulder: (363.28, 294.56) with score 0.979
+  - left_elbow: (514.37, 404.61) with score 0.990
+  - right_elbow: (353.58, 523.61) with score 0.957
+  - left_wrist: (607.64, 530.43) with score 0.985
+  - right_wrist: (246.78, 536.33) with score 0.950
+  - left_hip: (563.45, 577.89) with score 0.998
+  - right_hip: (544.08, 613.29) with score 0.997
+  - left_knee: (466.57, 862.51) with score 0.996
+  - right_knee: (518.49, 977.99) with score 0.996
+  - left_ankle: (691.56, 844.49) with score 0.960
+  - right_ankle: (671.32, 1100.90) with score 0.953
+Found person at [424.73594665527344, 68.82870757579803, 640.3419494628906, 492.8904126405716] with score 0.887
+  - nose: (840.26, 289.19) with score 0.991
+  - left_eye: (851.23, 259.92) with score 0.956
+  - right_eye: (823.10, 256.35) with score 0.955
+  - left_ear: (889.52, 278.10) with score 0.668
+  - right_ear: (799.80, 264.64) with score 0.771
+  - left_shoulder: (903.87, 398.65) with score 0.997
+  - right_shoulder: (743.88, 403.37) with score 0.988
+  - left_elbow: (921.63, 589.83) with score 0.989
+  - right_elbow: (699.56, 527.09) with score 0.934
+  - left_wrist: (959.21, 728.84) with score 0.984
+  - right_wrist: (790.88, 519.34) with score 0.945
+  - left_hip: (873.51, 720.07) with score 0.996
+  - right_hip: (762.29, 760.91) with score 0.990
+  - left_knee: (945.33, 841.65) with score 0.987
+  - right_knee: (813.06, 1072.57) with score 0.964
+  - left_ankle: (918.48, 1129.20) with score 0.871
+  - right_ankle: (886.91, 1053.95) with score 0.716
+```
+</details>