Images

class deepbench.image.ShapeImage(image_shape, object_noise_type='gaussian', object_noise_level=0.0)

Create an image that is a composition of multiple shapes

Parameters:

• image_shape (Tuple[int, int]) – Dimensions of the shape image.

• object_noise_type (str, optional) – Noise distribution applied to image. Defaults to “gaussian”.

• object_noise_level (float, optional) – Relative noise level (scale 0 to 1). Defaults to 0.0.

combine_objects(objects, object_params, instance_params=None, seed=42)

Utilize Image._generate_astro_objects to overlay all selected astro objects into one image If object parameters are not included in object list, defaults are used. Updates SkyImage.image.

Current input parameter assumptions (totally up to change): For a single image: [{

“object_type”:”<object_type>”, “object_parameters”:{<parameters for that object>}

}]

Parameters:

• objects (list) – str discriptors of the included object

• object_params (list) – Parameters of each object (ie, position in frame)

• seed (int, optional) – random seed for noise. Defaults to 42.

Returns:

image with objects and noise

Return type:

ndarray

class deepbench.image.SkyImage(image_shape, object_noise_level=0, object_noise_type='gaussian', scale=True)

combine_objects(objects, instance_params, object_params, seed=42)

Utilize Image._generate_astro_objects to overlay all selected astro objects into one image If object parameters are not included in object list, defaults are used. Updates SkyImage.image.

Current input parameter assumptions (totally up to change): For a single image: [{

“object_type”:”<object_type>”, “object_parameters”:{<parameters for that object>}

}]

Parameters:

• objects (list) – str discriptors of the included object

• instance_params (list) – Parameters for the instance of the object (ei, overall noise)

• object_params (list) – Parameters of each object (ei: position in frame)

• seed (int, optional) – random seed for noise. Defaults to 42.

Returns:

image with objects and noise

Return type:

ndarray