keras-team/keras · error · ValueError

Input images must have 3 channels, but received images with

Error message

Input images must have 3 channels, but received images with {channels} channels.

What it means

Raised by the image color-space conversion op (rgb_to_hsv / hsv_to_rgb) when the last dimension of `images` is not 3. The op requires RGB (or HSV) images with exactly 3 channels per pixel; grayscale (1 channel) or RGBA (4 channels) inputs are rejected.

Solutions

1. Convert grayscale to RGB by broadcasting the channel axis, and drop alpha: images = images[..., :3]
2. Pass data_format='channels_first' when your tensors are (N,C,H,W)

Example fix

# before
hsv = keras.ops.image.rgb_to_hsv(gray)  # gray.shape=(H,W,1)

# after
rgb = keras.ops.broadcast_to(gray, gray.shape[:-1] + (3,))
hsv = keras.ops.image.rgb_to_hsv(rgb)

c = images.shape[-1 if data_format == 'channels_last' else -3]
assert c is None or c == 3

def is_rgb(images, data_format='channels_last') -> bool:
    return images.shape[-1 if data_format == 'channels_last' else -3] in (3, None)

When it happens

Trigger: Passing grayscale (H,W,1) tensors or RGBA (H,W,4) tensors to rgb_to_hsv; passing channels_first data while leaving data_format at default so the wrong axis is read as channels.

Common situations: Applying HSV augmentation to a mixed dataset including grayscale images; alpha-channel PNGs; forgetting data_format='channels_first' for PyTorch-style tensors.

Related errors

• Invalid channel size: expected 3 (RGB) or 1 (Grayscale)…
• Invalid images dtype: expected float dtype. Received…
• Channels are allowed and the first and last dimensions.
• Expected `size` to be a tuple of 2 integers. Received: size=
• First dim of `coordinates` must be the same as the rank of…

AI-assisted analysis of keras-team/keras@7a34a03db6 (2026-08-25). Data as JSON: /api/errors/7b3413c409ab797a. Report an issue: GitHub.

Appendix: source

Thrown at keras/src/ops/image.py:108

    def compute_output_spec(self, images):
        images_shape = list(images.shape)
        dtype = images.dtype
        if len(images_shape) not in (3, 4):
            raise ValueError(
                "Invalid images rank: expected rank 3 (single image) "
                "or rank 4 (batch of images). "
                f"Received: images.shape={images_shape}"
            )
        if not backend.is_float_dtype(dtype):
            raise ValueError(
                "Invalid images dtype: expected float dtype. "
                f"Received: images.dtype={dtype}"
            )
        channels_axis = -1 if self.data_format == "channels_last" else -3
        channels = images_shape[channels_axis]
        if channels is not None and channels != 3:
            raise ValueError(
                "Input images must have 3 channels, but received images with "
                f"{channels} channels."
            )
        return KerasTensor(shape=images_shape, dtype=images.dtype)


@keras_export("keras.ops.image.rgb_to_hsv")
def rgb_to_hsv(images, data_format=None):
    """Convert RGB images to HSV.

    `images` must be of float dtype, and the output is only well defined if the
    values in `images` are in `[0, 1]`.

    All HSV values are in `[0, 1]`. A hue of `0` corresponds to pure red, `1/3`
    is pure green, and `2/3` is pure blue.

    Args:
        images: Input image or batch of images. Must be 3D or 4D.

View on GitHub (pinned to 7a34a03db6)