Identify images based on the annotated labels provided in the samples.
pip install -r requirements.txt
-
Image Processing
- Orientation Correction
- Image Binarization
- Denoising
- Deskewing
- Normalization
- Image Scaling
- Removing Horizontal and Vertical Lines
- Enchaining Contrast
- Image Segmentation through boxes
-
Text Processing
- Basic Text Cleaning (like removing repetitive special characters)
- Spelling Correction through SYMSPELL
- NER - For Names and Locations
- Keyword Extraction
- Profiling Text Data
-
Text Embeddings
- Transformers based Embeddings
- Sentence Transformers based Embeddings
- FastText Embeddings
-
Model Selection & Training
-
Model Evaluation
Some of the research findings are shared in the Document shared.