PyMuPDF delete only specific image same as 1.png from a pdf pages

[BeastHunterTest1]

hi i tried some code using chat gtb to delete all images same as 1.png from pdf called 1.pdf this is the code

`import fitz # PyMuPDF

def remove_image_from_pdf(input_pdf, output_pdf, image_name):
# Open the input PDF file
pdf_document = fitz.open(input_pdf)

for page_number in range(len(pdf_document)):
    page = pdf_document[page_number]
    xref_list = page.get_images(full=True)
    
    for xref in xref_list:
        base_image = pdf_document.extract_image(xref[0])
        image_data = base_image["image"]

        if image_data.endswith(b".png") and image_data.get("name") == image_name.encode():
            # Remove the image from the page
            page.apply_redact([xref])
            page.insert_redact(xref, width=0)
            page.apply_redactions()

# Save the modified PDF to the output file
pdf_document.save(output_pdf)
pdf_document.close()

if name == "main":
input_pdf_file = "1.pdf"
output_pdf_file = "1_remove.pdf"
image_name_to_remove = "1.png"

remove_image_from_pdf(input_pdf_file, output_pdf_file, image_name_to_remove)
print(f"Image '{image_name_to_remove}' removed and saved as '{output_pdf_file}'.")

`

and it...does save the new pdf but doesnot remove anything sooo i am trying to remove a translate img from pdf btw.
any help will be great. thanksss.

[JorjMcKie]

This a typical "Discussions" item, so I am going to move it there first.

[BeastHunterTest1]

hi sorry there was only req fection or bug option.

[JorjMcKie]

In PDF, images have no name! So you cannot delete them in this way.

This has nothing to do with PyMuPDF or MuPDF!

I recommend the following approach:
For a given image, compute the hash code (like MD5) from its binary content.
Then loop over the images and extract them via doc.extract_image(xref) as you did. For each image extracted in this way, compute the hash code from img["image"] and compare it with that of the file.
If equal, delete the image as indicated.

A lot of things may still go wrong - which is the reason why we won't make an official enhancement - e.g.:

• If the image file contains an alpha channel (transparency), there is no way to find its PDF equivalent using hash codes
• If the image file is black & white, then the PDF compression filter CCITTFaxDecode will probably have been used and the image extracted via doc.extract_image() will not deliver the same binary.
• ... probably more