<br />

|---|
| [![](https://www.tensorflow.org/images/GitHub-Mark-32px.png) View source on GitHub](https://github.com/tensorflow/models/blob/v2.20.0/official/vision/ops/preprocess_ops_3d.py#L201-L220) |

Decodes PNG or JPEG raw bytes string into a RGB uint8 Tensor.

    tfm.vision.preprocess_ops_3d.decode_image(
        image_string: tf.Tensor, channels: int = 0
    ) -> tf.Tensor

<br />

<br />

<br />

| ## Args ||
|---|---|
| `image_string` | A [`tf.Tensor`](https://www.tensorflow.org/api_docs/python/tf/Tensor) of type strings with the raw PNG or JPEG bytes where the first dimension is timesteps. |
| `channels` | Number of channels of the PNG image. Allowed values are 0, 1 and <br /> 1. If 0, the number of channels will be calculated at runtime and no static shape is set. |

<br />

<br />

<br />

| ## Returns ||
|---|---|
| A Tensor of shape \[T, H, W, C\] of type uint8 with the decoded images. ||

<br />