-
-
Contribute - We have provided an option to the user where they can either choose to upload a single image or in a bulk Recently Uploaded
-
Edit/Delete Tags – If the user is not satisfied with the AI generated tags, they have the option to add or delete the generated tags.
-
Spell Correction (Did you Mean)
-
Image extraction from pdfs
-
Phrase search from images
-
Similar Image band on tap
-
Upload video file
Inspiration ⚡
There are no existing solutions that can consolidate and re-use images. Reusing images results in cost savings and a streamlined workflow especially during this covid time when there has been a major transfer from traditional methods to virtual ones. Tagging and classifying the images reduces a lot of manual effort in searching for the required images. There are different scenarios where an image needs to be searched based on -
- Context – The images that we usually find in image search engines are usually labeled or tagged manually.
- Similarity – The current standards of the image search engines are not capable to extract features from the searched images and show recommended images based on image and tag similarity.
What it does 🤩
Project Aperture intends to provide an automated platform for the procedure of tagging and labeling the images by use of Image processing and Machine Learning to extract information from the images, using Convolutional Neural Networks and Natural Language Processing to generate labels and similar tags. It can extract and classify images from a group of images, documents as well as videos.
How we built it 🤖
Image tagging and labeling are done by use of Image processing and Machine Learning to extract information from the images, using CNN and NLP to generate labels and similar tags. The images are searched on the basis of labels rather than tags, to provide more accuracy. Azure Cognitive services are used for the Deep Learning API. The backend is built using Django and Django Rest Framework. The frontend is built using Django and Bootstrap.
Challenges we ran into 💪
While developing our project, we faced tons of challenges but a few noteworthy ones were: None of our teammates had experience working with Azure but we had to learn it from scratch! , extracting images from videos and pdf was an insurmountable feat, but we somehow managed to get it up and running after a lot of work. Uploading files in bulk was quite challenging, but with proper resources, we managed to pull this off as well.
Accomplishments that we're proud of 😎
We made a lot of features that we didn't even think would be possible to do and we had so much fun creating them. We were able to learn a lot during the short span of time and were able to build a fully working product. It feels great creating something which can actually help someone and make an impact on someone's life.
What we learned 🤓
We tried our hands on Python-based backend for the first time. None of our team members had previous experience working with Azure, We learnt how to use the Azure servies for the first time. We learnt how important database management is, while dealing with a large database.
What's next for Aperture 🔥
- Extending our platform such that the users can share images and talk in a private chat instead of using other social media platforms. (Data privacy)
- This platform may provide scope of collection of large amount of data from all around the world which will enable more innovative AI based solutions.
- Scope of using NLP generated tags in other languages instead of using only English.
Log in or sign up for Devpost to join the conversation.