This repository is for DATA Pacticum semester client project with Later, a social media and influencer marketing company. In this project, we wants to explore about if images be classified into useful categories at scale and at low cost, and with one million images with no words or labels, we'd like to conduct research on available open-source models and look for solutions to categorize images, as well as add labels and keywords to each piece of media.
Data: Access with northeastern google account: [https://drive.google.com/file/d/1GQDtwXT3Vkmi7-0wUKfW5JJoww3t3XfS/view?pli=1]
Useful resources:
- An intro of image recognition(Beginner friendly?): Pre Trained Models for Image Classification – PyTorch for Beginners [https://learnopencv.com/pytorch-for-beginners-image-classification-using-pre-trained-models/]
- A kaggle challenge of image recognition with 150,000 images to be recognized [https://www.kaggle.com/c/imagenet-object-localization-challenge/overview/description]
- Top 4 Pre-Trained Models for Image Classification with Python Code: [https://www.analyticsvidhya.com/blog/2020/08/top-4-pre-trained-models-for-image-classification-with-python-code/]
- Hugging face BRIA Background Removal [https://huggingface.co/briaai/RMBG-1.4]
- Clip Model [https://github.com/openai/CLIP/tree/main/clip]
- Intro to image segmentation [https://huggingface.co/tasks/image-segmentation]
- What is an image embedding? [https://blog.roboflow.com/what-is-an-image-embedding/]
- How to Detect Segments in Videos with Computer Vision [https://blog.roboflow.com/detect-video-segments/]
- VILA: On Pre-training for Visual Language Models [https://vila.mit.edu/]