diff --git a/image_prediction/image_extractor/extractors/parsable.py b/image_prediction/image_extractor/extractors/parsable.py
index 885cbae..d115059 100644
--- a/image_prediction/image_extractor/extractors/parsable.py
+++ b/image_prediction/image_extractor/extractors/parsable.py
@@ -1,22 +1,18 @@
 import atexit
-import json
-import traceback
 from _operator import itemgetter
 from functools import partial, lru_cache
-from itertools import chain, starmap, filterfalse
+from itertools import chain, filterfalse
 from operator import itemgetter
-from typing import Iterable, Iterator, List, Union
+from typing import List, Union
 
-import IPython
 import fitz
 import numpy as np
 from PIL import Image
-from funcy import merge, compose, rcompose, keep, identity
+from funcy import merge, compose, rcompose, keep
 from pymonad.maybe import Maybe, Nothing, Just
 from pymonad.tools import curry
 
 from image_prediction.exceptions import InvalidBox, BadXref
-from image_prediction.formatter.formatters.enum import EnumFormatter
 from image_prediction.image_extractor.extractor import ImageExtractor, ImageMetadataPair
 from image_prediction.info import Info
 from image_prediction.stitching.stitching import stitch_pairs
@@ -25,8 +21,6 @@ from image_prediction.utils import get_logger
 
 logger = get_logger()
 
-Maybe
-
 
 class ParsablePDFImageExtractor(ImageExtractor):
     def __init__(self, verbose=False, tolerance=0):
@@ -57,29 +51,10 @@ class ParsablePDFImageExtractor(ImageExtractor):
         image_metadata_pairs = [pair.value for pair in maybe_image_metadata_pairs if pair.is_just()]
         clear_caches()
 
-        #  TODO: In the future, consider to introduce an image validator as a pipeline component rather than doing the
-        #   validation here. Invalid images can then be split into a different stream and joined with the intact images
-        #   again for the formatting step.
-        image_metadata_pairs = self.__filter_valid_images(image_metadata_pairs)
-
         image_metadata_pairs = stitch_pairs(list(image_metadata_pairs), tolerance=self.tolerance)
 
         yield from image_metadata_pairs
 
-    @staticmethod
-    def __filter_valid_images(image_metadata_pairs: Iterable[ImageMetadataPair]) -> Iterator[ImageMetadataPair]:
-        def validate(image: Image.Image, metadatum: dict):
-            try:
-                # TODO: stand-in heuristic for testing if image is valid => find cleaner solution (RED-5148)
-                image.resize((100, 100)).convert("RGB")
-                return ImageMetadataPair(image, metadatum)
-            except (OSError, Exception):
-                metadatum = json.dumps(EnumFormatter()(metadatum), indent=2)
-                logger.warning(f"Invalid image encountered. Image metadata:\n{metadatum}\n\n{traceback.format_exc()}")
-                return None
-
-        return keep(starmap(validate, image_metadata_pairs))
-
 
 def extract_pages(doc, page_range):
     page_range = range(page_range.start + 1, page_range.stop + 1)
@@ -88,6 +63,16 @@ def extract_pages(doc, page_range):
     yield from pages
 
 
+def validate_image(image: Image.Image) -> Maybe:
+    try:
+        # TODO: stand-in heuristic for testing if image is valid => find cleaner solution (RED-5148)
+        image.resize((100, 100)).convert("RGB")
+        return Just(image)
+    except (OSError, Exception):
+        logger.warning(f"Invalid image encountered.")
+        return Nothing
+
+
 def extract_valid_metadata(doc: fitz.fitz.Document, page: fitz.fitz.Page):
     return compose(
         list,
@@ -98,8 +83,9 @@ def extract_valid_metadata(doc: fitz.fitz.Document, page: fitz.fitz.Page):
 
 
 def metadatum_to_image_metadata_pair(doc, metadatum: dict) -> Maybe:
-    maybe_image = xref_to_maybe_image(doc, metadatum[Info.XREF])
-    return make_maybe_image_metadata_pair(maybe_image, Just(metadatum))
+    maybe_image = xref_to_maybe_image(doc, metadatum[Info.XREF]).bind(validate_image)
+    maybe_image_metadata_pair = make_maybe_image_metadata_pair(maybe_image, Just(metadatum))
+    return maybe_image_metadata_pair
 
 
 def add_alpha_channel_info(doc, metadata):