ChatGPT Advanced Data Analysis Hack: Extract Text From Images (OCR)

[@akratlapidus2390]

Man, you are amazing! I liked your videos about autonomous agents. Then you brought me the idea of creating my own Jarvis assistant. You are a true reference figure in the AI world. Thanks a lot! 😁

[@DJPapzin]

🎯 Key Takeaways for quick navigation: 00:00 ⌨️ *Introduction to OCR Text Extraction* - OCR (Optical Character Recognition) can be used to extract text from images or PDFs. - The video sets the stage for the demonstration by introducing the need for extracting text from images. - The presenter mentions the tools (like the Snip tool) and the images to be used for the demonstration. 01:09 🛠️ *Setting Up and Prompt Overview* - The presenter prepares to use the Code Interpreter and mentions the system prompts used. - Introduces the task of extracting text from images using OCR. - Shares the prompt instructing to upload images in a zip file and use OCR to extract text, followed by summarizing and saving it to a file. 02:34 📚 *Explanation of OCR and Required Libraries* - Briefly explains OCR (Optical Character Recognition) and its role in extracting text from images. - Mentions the Python library used for OCR and directs to the description for the library link. - Emphasizes the importance of having the required modules installed for the Code Interpreter task. 03:05 ⚙️ *Running the Code Interpreter Task* - Describes the step-by-step plan for the Code Interpreter task: Unzipping files, extracting images, summarizing text, and writing to a file. - Demonstrates the successful execution of unzipping the files. - Shares the output, highlighting the extracted text from each image, and mentions the summary file. 04:42 🚀 *Conclusion and Future Prompts* - Concludes the demonstration and highlights the ease of using OCR for text extraction. - Encourages viewers to try out the provided prompts on the presenter's website. - Teases the future upload of more interesting prompts on the website. Made with HARPA AI

[@BirgittaGranstrom]

Yeaaa! Please keep those “tutorials” coming! It takes a brilliant brain to figure out practical uses for this fascinating Code Interpreter. I almost have the feeling of that this is too good to be true;-)

[@maddercat]

I tried with a pdf, doesn't seem to work I've seen others get it to read a pdf?

[@sdtimeless]

Do we need to install anything for ‘pytesseract==0.3.8*’ to work as shown?

[@1986xuan]

My wish for the code interpreter to do is not only able to extract text but also tables in PDFs.. I noticed the code interpreter struggles with getting the right information from columns and rows in a table

[@PRINTHINK]

I experienced the same

[@VIthingsforsaleortrade]

Nice video. I just tried it and it works but I just told it to use OCR to extract and summarize the text from this image and it did

[@pecetowiec5851]

hi, it is possible to grab text from video? Fg. i record video of my gameplay , and scroll throu in game marketplace. I need to have this all items and prices (without duplicats) in text file or database. It is possible with AI?

[@The_Outkeeper]

thank you for the prompt. great work. I learned a lot through your videos. I just got shut down from it though. The text extraction process is still timing out, even when processing a single page at a time. This could be due to the high resolution of the images, the complexity of the document layout, or a large amount of text on each page. All of these factors make OCR a computationally expensive operation. interesting though.

[@bierzudir6713]

cool stuff man" THANKS A MILLION

[@micbab-vg2mu]

great video!

[@edersolis5097]

Love it! I’m wondering if this can be applied to images with filters or light room presets. Would it be able to tell you the presets or the filter being used?

[@FilmFactry]

Can it save to a Google drive or docs? It would be useful to always save your work, and maybe GPT can go back and read from the drive or doc at a later time. Great video!

[@laugedyret]

Why use "ignore all previous instructions" in a new prompt?

[@WhySoBroke]

I don’t see the point of your long winded prompt, I exported a pdf into images, zipped and then just asked chatgp CI for unzipping, converting pics to text, I even just simply asked to convert any tables into excel and provide output files including a summary. I literally typed it exact like that and it executed all the instructions. I don’t think there is much of a need of all the prompt entering stuff you are doing

[@fabriziocasula]

wow!! thanks

[@jesseburstrom5920]

I was thrown into Python project black box testing. Have some 70 functions 2000 sub lines of code and used AI all the way building it. Now Code Interpreter - I laugh why am I here?!

[@socialnet5795]

Hñ

[@divisarvaraidu41]

Windows Power toys is much powerful than this ig

[@technolus5742]

Not really, they do different things. The ability to summarize and directly answer questions about your data is not really part of power toys.