DeepDive2Img

• 

  Home Page

• 

  Categorized an image

• 

  Categorized another image

Inspiration

The sentiment analysis of tweets and Text Extraction served as our source of inspiration. We believed that we could create a mixture of both and convert them to something very fruitful using Cohere.

What it does

The projects will extract the text from the image which helps full when used in a social media platform to link the text of the image with the image to identify the post ex. while searching. and with that it classifies the categories the image according to the emotion of the text contained in the image which will help us to categories the images which can be helpful for suggestions of the post according to the interest of the users.

How we built it

For the Text extraction from the image we used computer vision along with EasyOCR library with Python and We used Django to build our website and the backend is hence running on django(Python) We used cohere:AI online NLP platform to build a NLP model which will classify the extracted text from a image according to emotions. we used regex feature form python for preprocessing of the data. We have used a Tweet's dataset which contains Messages of different types of emotions.

Challenges we ran into

To arrange the functionality in a easy and user-friendly way as the working is more for backend part, but we did it with the help of django

Accomplishments that we're proud of

we are able to satisfy the aim of our topic and with different text example we can prove it. also we were able to represent the functionality in a user friendly way using a website

What we learned

Using Regexlib, CohereAI, OCR . Also sentiment analysis

What's next for DeepDive2Img

We can identify the objects as well from the image and then can collect the information form aa source about that objects dn can link that to image and categories the image more generally

Built With

• cohereai
• computervision
• django
• ocr
• python