JPG Process

get_orientation_code(stream: str | Path | bytes) -> ROTATE | None

jpgencode(img: np.ndarray, quality: int = 90) -> bytes | None

jpgdecode(byte_: bytes) -> np.ndarray | None

jpgread(img_file: str | Path) -> np.ndarray | None

These helpers provide JPEG encode/decode/read support and apply EXIF orientation during decode.

Description

• get_orientation_code: Reads EXIF orientation and returns the corresponding rotate code (used by jpgdecode / jpgread).
• jpgencode: Encodes a numpy image into JPEG bytes; returns None on failure.
• jpgdecode: Decodes JPEG bytes into a numpy image and applies EXIF orientation; returns None on failure.
• jpgread: Reads a JPEG file and returns a numpy image (via jpgdecode).

Dependencies

• This module uses turbojpeg (PyTurboJPEG) for JPEG encode/decode.

Parameters

jpgencode

• img (np.ndarray): Image array.
• quality (int): JPEG quality (1 to 100). Default is 90.

jpgdecode

• byte_ (bytes): JPEG bytes.

jpgread

• img_file (str | Path): JPEG file path.

Examples

jpgencode

from capybara.vision.improc import imread, jpgencode

img = imread('lena.png')
encoded_bytes = jpgencode(img, quality=95)

jpgdecode

from capybara.vision.improc import jpgdecode

decoded_img = jpgdecode(encoded_bytes)

jpgread

from capybara.vision.improc import jpgread

img_array = jpgread('path/to/image.jpg')