top of page

AI based OCR, an essential patent invalidity tool – A case study

With the emergence of technological advancements, one term is ubiquitous, Artificial Intelligence (AI). Across various fields, AI is changing how we look at current technologies. One technology transformed with the implementation of AI is OCR, Optical Character Recognition. OCR processes images of text and converts that text into machine-readable forms. In other words, it takes handwritten or typed text within physical documents and converts them into digital formats.


In the 1990s, many people used OCR, sometimes called text recognition, to convert physical documents into digital files. Since then, the quality of OCR technology has improved, but demand has increased for broader usability. Recent developments with AI have amplified OCR’s utility thanks to higher accuracy and greater speed. With the benefit of AI, human supervision isn’t needed at every step. As an example, let’s say an OCR tool was converting print bills into digital copies. Let’s say the OCR scanner identified the bill total as $100, when it was really $1,000. Without AI, the OCR tool wouldn’t pick up on this mistake and it would take manual effort to scrutinize it. With AI tools, however, it reviews the entire document and fix the error without a human needing to supervise.

Patent Invalidation or Opposition Search is conducted to check the validity of a patent. Mostly it is conducted to invalidate a patent in a patent infringement lawsuit; or when a patent application for a similar invention is filed by another party. Invalidity search is focused on the claims and file history of the patent to be invalidated. In addition to revoking a patent/application, invalidity searches help perform IP due diligence. Mostly the invalidity searches conducted by search experts rely on the textual description available in patent specifications (such as abstract, claims or description) or non-patent literatures (such as journals, research papers, thesis). For structural novelty aspects of an invention, images are also reviewed. AI based OCR tools are helpful for cases where novelty or inventive aspect is disclosed as text in the images or when a non-patent literature is present in an image format pdf.


In a recent patent invalidity (Bool’s Eye) search, we were assigned to look for a particular ingredient in the composition of a toothpaste. Our team tried all combinations of search strategy from narrow to broad and was exhausted with everything they could do on paid databases for patent or non-patent literatures. We then shifted our focus to check for images in patents and non-patent images showing the composition of toothpaste. It was a very tedious task for our team to check every image thoroughly. However, our AI based OCR tool came very handy wherein we uploaded all the images of the patents as well as non-patents and in no time, it made the text in the images recognizable. We were successful in identifying one patent as well one toothpaste brand image from 1960s that disclosed the required ingredient in the toothpaste composition. The document was a bang-on killer art and it helped our client prepare a strong attack against the target patent. When we checked why we couldn’t identify the document in our search strategy earlier, we found that the document wasn’t captured in text-based strategy. Surprisingly, the required ingredient wasn’t disclosed anywhere in the full patent specification. Instead, the required ingredient was disclosed as a substitute to another second ingredient, that disclosed everywhere in the patent specification, in a composition chart shown as image. The non-patent image clearly mentioned the required ingredient at the back of the toothpaste package.


This learning experience of ours strengthens our trust in our team’s expertise and technology (specially AI). We have now made it as a standard practice in all our invalidity searches to use OCR tools to look for evidence in images or pdfs. Though it must be noted that we still do not completely rely on AI based tools. Manual searching (using Boolean operators etc.) by our search experts is still at the core of our competency, and we use AI tools only to complement the effort by our experts.


If you need a patent invalidity expert who not only understands the technology, industry, and your specific needs, but also stay with you on each step during assignment and make sure you get the results crucial for your win, then you are at the right place. Our experts, at Boolean IP, have a total experience of conducting 500+ successful patent invalidity searches. Clubbing the experience and expertise of our techno-legal experts, and AI-powered tools, Boolean IP provides high quality IP solutions that help you win. Get in touch with your queries.

Recent Posts

See All
bottom of page