svkcore.visualize

svkcore.visualize.visualize

svkcore.visualize.visualize.cv2image2pil(cv2_image: ndarray) → Image

Convert openCV format image to PIL.Image.Image

Parameters:

cv2_image – openCV format image instance

Returns:

converted Image.Image instance

svkcore.visualize.visualize.draw_annotation(image, annotation, name2cls, color_table=None, add_unknown_name=False)

Draw DTAnnotation to an image

Parameters:

• image (PIL.Image.Image) – A PIL.Image.Image object

• annotation (DTAnnotation) – An instance of DTAnnotation

• name2cls (dict) – a dict of name to its class id number

• color_table (list) – each class colors

• add_unknown_name (bool) – whether add a new name to name2cls, default is False

Returns:

drew image

Return type:

PIL.Image.Image

svkcore.visualize.visualize.draw_boxes(image, boxes, color='red', width=0, fullfill=False)

Draw boxes on image

Parameters:

• image – A PIL.Image object

• boxes – a list of boxes

• color – boxes color

• width – line width

• fullfill – full fill boxes or not

Returns:

drew image

svkcore.visualize.visualize.draw_boxes_texts(image, boxes, texts, width=1, color='red')

Draw boxes and its text information on image

Parameters:

• image – A PIL.Image object

• boxes – A list of box

• texts – A list of text

• width – Line width, determines thickness of box and font size of text

• color – Color of boxes and texts

Returns:

drew image

svkcore.visualize.visualize.draw_detection_result(image, boxes, classes, display_strings, color_nums=100, scale=-1.0)

Draw detection result

Parameters:

• image – A PIL.Image object

• boxes – A list of box

• classes – A list of detection class index

• display_strings – A list of display string

• color_nums – The max number of different colors

• scale – visualize box and text scale. Default -1.0 means auto adjust scale by input image

Returns:

drew image

svkcore.visualize.visualize.draw_lines(image, lines, color='red', width=0)

Draw boxes on lines

Parameters:

• image – A PIL.Image object

• lines – a list of lines

• color – lines color

• width – line width

Returns:

drew image

svkcore.visualize.visualize.draw_masks(image, masks, color='red', alpha=0.5)

Draw masks on image

Parameters:

• image – A PIL.Image object

• masks – a list of masks

• color – mask color

• alpha – transport alpha

Returns:

drew image

svkcore.visualize.visualize.draw_points(image, points, color='red', scale=3, shape='.')

Draw boxes on points

Parameters:

• image – A PIL.Image object

• points – a list of points

• color – points color

• scale – points scale

• shape – visualize shape, ‘t’ for triangle else for circle

Returns:

drew image

svkcore.visualize.visualize.draw_polygons(image, polygons, color='red', width=0, fullfill=False)

Draw polygons on image

Parameters:

• image – A PIL.Image object

• polygons – a list of polygons

• color – mask color

• width – line width

• fullfill – full fill polygon or not

Returns:

drew image

svkcore.visualize.visualize.draw_texts(image, xys, texts, color='red', back_color=None, font_size=12, position='topleft', offset=(0, 0), margin=(0, 0, 0, 0))

Draw texts on image

Parameters:

• image – A PIL.Image object

• xys – A list of corner’s coordinate where text align

• texts – A list of texts

• color – Text color

• back_color –

• font_size – Font size

• position – text align position, enum string like: topleft/topright/bottomleft/bottomright/manu

• offset – A tuple vector of offset for ‘manu’ position

• margin – A tuple (left, top, right, bottom) denotes the margins to back boarder

Returns:

drew image

svkcore.visualize.visualize.generate_colors(num)

Generate colors for drawing bounding boxes

svkcore.visualize.visualize.images_gallery(image_list: Tuple[Image] | List[Image], n_cols: int = 6, n_rows: int | None = None, cell_size: Tuple[int] | List[int] = (224, 224), pad: int = 16, align: int = 0, back_color: Tuple[int] | List[int] | str = 'black', same_scale: bool = False) → Image

Paste a list of images into one panel for better visualize

Parameters:

• image_list – A list of Image.Image instance.

• n_cols – Max number of images to show in each row. If the number of image_list is less than n_cols, n_cols will be set as the number of image_list. Default value is 6.

• n_rows –

  Number of rows to show images. If n_cols is set, n_rows will be automatically calculated by

  \[ceil(\frac{N}{n\_cols})\]

  Otherwise n_cols will be automatically calculated. Default value is None.

• cell_size – The size of cell where each image is placed in.

• pad – The pad width/height between two cells.

• align – The align mode. Set 0(center), 1(left/up), 2(right/bottom) to choose align mode in each cell.

• back_color – Background color.

• same_scale – A boolean value indicates whether to use a same scale factor to resize all images in image_list. Set it to be True if you want to visualize the sizes of different images.

Returns:

Pasted Image.Image instance.

svkcore.visualize.visualize.pil2cv2image(image: Image) → ndarray

Convert PIL.Image.Image to openCV format image

Parameters:

image – an instance of PIL.Image.Image

Returns:

converted openCV format image