Retrieval-Augmented Generation (RAG) has become a popular technique for enhancing the reliability and utility of Large Language Models (LLMs) by grounding responses in external documents. Traditional RAG systems rely on Optical Character Recognition (OCR) to first process scanned documents into text. However, even state-of-the-art OCRs can introduce errors, especially in degraded or complex documents. Recent vision-language approaches, such as ColPali, propose direct visual embedding of documents, eliminating the need for OCR. This study presents a systematic comparison between a vision-based RAG system (ColPali) and more traditional OCR-based pipelines utilizing Llama 3.2 (90B) and Nougat OCR across varying document qualities. Beyond conventional retrieval accuracy metrics, we introduce a semantic answer evaluation benchmark to assess end-to-end question-answering performance. Our findings indicate that while vision-based RAG performs well on documents it has been fine-tuned on, OCR-based RAG is better able to generalize to unseen documents of varying quality. We highlight the key trade-offs between computational efficiency and semantic accuracy, offering practical guidance for RAG practitioners in selecting between OCR-dependent and vision-based document retrieval systems in production environments.

Although convolutional neural networks (CNNs) are inspired by the mechanisms behind human visual systems, they diverge on many measures such as ambiguity or hardness. In this paper, we make a surprising discovery: there exists a (nearly) universal score function for CNNs whose correlation is statistically significant than the widely used model confidence with human visual hardness. We term this function as angular visual hardness (AVH) which is given by the normalized angular distance between a feature embedding and the classifier weights of the corresponding target category in a CNN. We conduct an in-depth scientific study. We observe that CNN models with the highest accuracy also have the best AVH scores. This agrees with an earlier finding that state-of-art models tend to improve on the classification of harder training examples. We find that AVH displays interesting dynamics during training: it quickly reaches a plateau even though the training loss keeps improving. This suggests the need for designing better loss functions that can target harder examples more effectively. Finally, we empirically show significant improvement in performance by using AVH as a measure of hardness in self-training methods for domain adaptation.

We present a study which is concerned with word recognition rates for heavily degraded documents. We compare human with machine reading capabilities in a series of experiments, which explores the interaction of word/non-word recognition, word frequency and legality of non-words with degradation level. We also study the influence of character segmentation, and compare human performance with that of our artificial neural network model for reading. We found that the proposed computer model uses word context as efficiently as humans, but performs slightly worse on the pure character recognition task. 1 Introduction Optical Character Recognition (OCR) of machine-print document images ·has matured considerably during the last decade. Recognition rates as high as 99.5% have been reported on good quality documents. However, for lower image resolutions (200 Dpl and below), noisy images, images with blur or skew, the recognition rate declines considerably. In bad quality documents, character segmentation is as big a problem as the actual character recognition.

We present a study which is concerned with word recognition rates for heavily degraded documents. We compare human with machine reading capabilitiesin a series of experiments, which explores the interaction of word/non-word recognition, word frequency and legality of non-words with degradation level. We also study the influence of character segmentation, andcompare human performance with that of our artificial neural network model for reading. We found that the proposed computer model uses word context as efficiently as humans, but performs slightly worse on the pure character recognition task. 1 Introduction Optical Character Recognition (OCR) of machine-print document images ·has matured considerably during the last decade. Recognition rates as high as 99.5% have been reported ongood quality documents. However, for lower image resolutions (200 Dpl and below), noisy images, images with blur or skew, the recognition rate declines considerably. Inbad quality documents, character segmentation is as big a problem as the actual character recognition.