Extract Images from Presentation Shapes in Python

Extract Images from Shapes

Images are often added to shapes and also frequently used as slides' backgrounds. The image objects are added through IImageCollection, which is a collection of IPPImage objects.

This article explains how you can extract the images added to presentations.

To extract an image from a presentation, you have to locate the image first by going through every slide and then going through every shape. Once the image is found or identified, you can extract it and save it as a new file. XXX

import aspose.slides as slides

def get_image_format(image_type):
    return {
        "jpeg": slides.ImageFormat.JPEG,
        "emf": slides.ImageFormat.EMF,
        "bmp": slides.ImageFormat.BMP,
        "png": slides.ImageFormat.PNG,
        "wmf": slides.ImageFormat.WMF,
        "gif": slides.ImageFormat.GIF,
    }.get(image_type, slides.ImageFormat.JPEG)

with slides.Presentation("pres.pptx") as pres:
    #Accessing the presentation
    
    slideIndex = 0
    image_type = ""
    ifImageFound = False
    for slide in pres.slides:
        slideIndex += 1
        #Accessing the first slide
        image_format = slides.ImageFormat.JPEG

        back_image = None
        file_name = "BackImage_Slide_{0}{1}.{2}"
        is_layout = False

        if slide.background.fill_format.fill_type == slides.FillType.PICTURE:
            #Getting the back picture  
            back_image = slide.background.fill_format.picture_fill_format.picture.image
        elif slide.layout_slide.background.fill_format.fill_type == slides.FillType.PICTURE:
            #Getting the back picture  
            back_image = slide.layout_slide.background.fill_format.picture_fill_format.picture.image
            is_layout = True

        if back_image is not None:
            #Setting the desired picture format 
            image_type = back_image.content_type.split("/")[1]
            image_format = get_image_format(image_type)

            back_image.image.save(
                file_name.format("LayoutSlide_" if is_layout else "", slideIndex, image_type), 
                image_format)

        for i in range(len(slide.shapes)):
            shape = slide.shapes[i]
            shape_image = None

            if type(shape) is slides.AutoShape and shape.fill_format.fill_type == slides.FillType.PICTURE:
                shape_image = shape.fill_format.picture_fill_format.picture.image
            elif type(shape) is slides.PictureFrame:
                shape_image = shape.picture_format.picture.image

            if shape_image is not None:
                image_type = shape_image.content_type.split("/")[1]
                image_format = get_image_format(image_type)

                shape_image.image.save(
                                file_name.format("shape_"+str(i)+"_", slideIndex, image_type), 
                                image_format)

FAQ

Can I extract the original image without any cropping, effects, or shape transformations?

Yes. When you access a shape’s image, you get the image object from the presentation’s image collection, meaning the original pixels without cropping or styling effects. The workflow goes through the presentation’s image collection and PPImage objects, which store the raw data.

Is there a risk of duplicating identical files when saving many images at once?

Yes, if you save everything indiscriminately. A presentation’s image collection can contain identical binary data referenced by different shapes or slides. To avoid duplicates, compare hashes, sizes, or contents of the extracted data before writing.

How can I determine which shapes are linked to a specific image from the presentation’s collection?

Aspose.Slides does not store reverse links from PPImage to shapes. Build a mapping manually during traversal: whenever you find a reference to an PPImage, record which shapes use it.

Can I extract images embedded inside OLE objects, such as attached documents?

Not directly, because an OLE object is a container. You need to extract the OLE package itself and then analyze its contents using separate tools. Presentation picture shapes work via PPImage; OLE is a different object type.