Charm-tokenizer 1.0.10

Official implementation of Charm tokenizer for ViTs

Charm: The Missing Piece in ViT fine-tuning for Image Aesthetic Assessment

Accepted at CVPR 2025

We introduce Charm , a novel tokenization approach that preserves Composition, High-resolution, Aspect Ratio, and Multi-scale information simultaneously. By preserving critical aesthetic information, Charm achieves significant performance improvement across different image aesthetic and quality assessment datasets.

Quick Inference

• Step 1) Check our GitHub Page and install the requirements.

pip install -r requirements.txt

---

• Step 2) Install Charm tokenizer.

pip install Charm-tokenizer

---

• Step 3) Tokenization + Position embedding preparation

from Charm_tokenizer.ImageProcessor import Charm_Tokenizer

img_path = r"img.png"

charm_tokenizer = Charm_Tokenizer(patch_selection='frequency', training_dataset='tad66k', without_pad_or_dropping=True)
tokens, pos_embed, mask_token = charm_tokenizer.preprocess(img_path)

Charm Tokenizer has the following input args:

• patch_selection (str): The method for selecting important patches
  • Options: 'saliency', 'random', 'frequency', 'gradient', 'entropy', 'original'.
• training_dataset (str): Sets the number of ViT input tokens to match a specific training dataset from the paper.
  • Aesthetic assessment datasets: 'aadb', 'tad66k', 'para', 'baid'.
  • Quality assessment datasets: 'spaq', 'koniq10k'.
• backbone (str): The ViT backbone model (default: 'facebook/dinov2-small').
• factor (float): The downscaling factor for less important patches (default: 0.5).
• scales (int): The number of scales used for multiscale processing (default: 2).
• random_crop_size (tuple): Used for the 'original' patch selection strategy (default: (224, 224)).
• downscale_shortest_edge (int): Used for the 'original' patch selection strategy (default: 256).
• without_pad_or_dropping (bool): Whether to avoid padding or dropping patches (default: True).

The output is the preprocessed tokens, their corresponding positional embeddings, and a mask token that indicates which patches are in high resolution and which are in low resolution.

---

• Step 4) Predicting aesthetic/quality score

from Charm_tokenizer.Backbone import backbone

model = backbone(training_dataset='tad66k', device='cpu')
prediction = model.predict(tokens, pos_embed, mask_token)

Note:

1. While random patch selection during training helps avoid overfitting,for consistent results during inference, fully deterministic patch selection approaches should be used.
2. For the training code, check our GitHub Page.