RTMO-t

File size: 5,535 Bytes

---
library_name: transformers.js
tags:
- pose-estimation
license: apache-2.0
---


https://github.com/open-mmlab/mmpose/tree/main/projects/rtmo 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/RTMO-t`.

```js
import { AutoModel, AutoProcessor, RawImage } from '@xenova/transformers';

// Load model and processor
const model_id = 'Xenova/RTMO-t';
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, original_sizes, reshaped_input_sizes } = await processor(image);

// Predict bounding boxes and keypoints
const { dets, keypoints } = await model({ input: pixel_values });

// Select the first image
const predicted_boxes = dets.tolist()[0];
const predicted_points = keypoints.tolist()[0];
const [height, width] = original_sizes[0];
const [resized_height, resized_width] = reshaped_input_sizes[0];

// Compute scale values
const xScale = width / resized_width;
const yScale = height / resized_height;

// Define thresholds
const point_threshold = 0.3;
const box_threshold = 0.3;

// Display results
for (let i = 0; i < predicted_boxes.length; ++i) {
    const [xmin, ymin, xmax, ymax, box_score] = predicted_boxes[i];
    if (box_score < box_threshold) continue;

    const x1 = (xmin * xScale).toFixed(2);
    const y1 = (ymin * yScale).toFixed(2);
    const x2 = (xmax * xScale).toFixed(2);
    const y2 = (ymax * yScale).toFixed(2);

    console.log(`Found person at [${x1}, ${y1}, ${x2}, ${y2}] with score ${box_score.toFixed(3)}`)
    const points = predicted_points[i]; // of shape [17, 3]
    for (let id = 0; id < points.length; ++id) {
        const label = model.config.id2label[id];
        const [x, y, point_score] = points[id];
        if (point_score < point_threshold) continue;
        console.log(`  - ${label}: (${(x * xScale).toFixed(2)}, ${(y * yScale).toFixed(2)}) with score ${point_score.toFixed(3)}`);
    }
}
```

<details>

<summary>See example output</summary>

```
Found person at [411.10, 63.87, 647.68, 505.40] with score 0.986
  - nose: (526.09, 119.83) with score 0.874
  - left_eye: (539.01, 110.39) with score 0.696
  - right_eye: (512.50, 111.08) with score 0.662
  - left_shoulder: (563.59, 171.10) with score 0.999
  - right_shoulder: (467.38, 160.82) with score 0.999
  - left_elbow: (572.72, 240.61) with score 0.999
  - right_elbow: (437.86, 218.20) with score 0.998
  - left_wrist: (603.74, 303.53) with score 0.995
  - right_wrist: (506.01, 218.68) with score 0.992
  - left_hip: (536.00, 306.25) with score 1.000
  - right_hip: (472.79, 311.69) with score 0.999
  - left_knee: (580.82, 366.38) with score 0.996
  - right_knee: (500.25, 449.72) with score 0.954
  - left_ankle: (572.21, 449.52) with score 0.993
  - right_ankle: (541.37, 436.71) with score 0.916
Found person at [93.58, 19.64, 492.62, 522.45] with score 0.909
  - left_shoulder: (233.76, 109.57) with score 0.971
  - right_shoulder: (229.56, 100.34) with score 0.950
  - left_elbow: (317.31, 162.73) with score 0.950
  - right_elbow: (229.98, 179.31) with score 0.934
  - left_wrist: (385.59, 219.03) with score 0.870
  - right_wrist: (161.31, 230.74) with score 0.952
  - left_hip: (351.23, 243.42) with score 0.998
  - right_hip: (361.94, 240.70) with score 0.999
  - left_knee: (297.77, 382.00) with score 0.998
  - right_knee: (306.07, 393.59) with score 1.000
  - left_ankle: (413.48, 354.16) with score 1.000
  - right_ankle: (445.30, 488.11) with score 0.999
Found person at [-1.46, 50.68, 160.66, 371.74] with score 0.780
  - nose: (80.17, 81.16) with score 0.570
  - left_eye: (85.17, 75.45) with score 0.383
  - right_eye: (70.20, 77.09) with score 0.382
  - left_shoulder: (121.30, 114.98) with score 0.981
  - right_shoulder: (46.56, 114.41) with score 0.981
  - left_elbow: (144.09, 163.76) with score 0.777
  - right_elbow: (29.69, 159.24) with score 0.886
  - left_wrist: (142.31, 205.64) with score 0.725
  - right_wrist: (6.24, 199.62) with score 0.876
  - left_hip: (108.07, 208.90) with score 0.992
  - right_hip: (64.72, 212.01) with score 0.996
  - left_knee: (115.26, 276.52) with score 0.998
  - right_knee: (65.09, 283.25) with score 0.998
  - left_ankle: (126.09, 340.42) with score 0.991
  - right_ankle: (63.88, 348.88) with score 0.977
Found person at [526.35, 36.25, 650.42, 280.90] with score 0.328
  - nose: (554.06, 71.87) with score 0.901
  - left_eye: (562.10, 66.30) with score 0.928
  - right_eye: (546.65, 66.36) with score 0.746
  - left_ear: (575.98, 68.17) with score 0.658
  - left_shoulder: (588.04, 102.61) with score 0.999
  - right_shoulder: (526.00, 102.94) with score 0.704
  - left_elbow: (618.11, 149.18) with score 0.984
  - left_wrist: (630.77, 189.42) with score 0.961
  - left_hip: (578.74, 181.42) with score 0.966
  - right_hip: (530.33, 176.46) with score 0.698
  - left_knee: (568.74, 233.01) with score 0.958
  - right_knee: (542.44, 243.87) with score 0.687
  - left_ankle: (585.17, 284.79) with score 0.838
  - right_ankle: (550.07, 292.19) with score 0.435
```

</details>