Grounding Descriptions in Images informs Zero-Shot Visual Recognition

December 05, 2024 ยท Entered Twilight ยท ๐Ÿ› arXiv.org

๐Ÿ’ค TWILIGHT: Eternal Rest
Repo abandoned since publication

Repo contents: .gitignore, LICENSE, README.md, assets, box_supervision.py, bpe_simple_vocab_16e6.txt.gz, clip, configs, dataloaders, estimate_queries.py, evaluate.py, gpt4_anns.py, model, our_eval.py, tokenizer.py, train.py, train_clip.py, utils, vqloc_drop.yml

Authors Shaunak Halbe, Junjiao Tian, K J Joseph, James Seale Smith, Katherine Stevo, Vineeth N Balasubramanian, Zsolt Kira arXiv ID 2412.04429 Category cs.CV: Computer Vision Cross-listed cs.LG Citations 1 Venue arXiv.org Repository https://github.com/shaunak27/grain-clip โญ 10 Last Checked 3 months ago
Abstract
Vision-language models (VLMs) like CLIP have been cherished for their ability to perform zero-shot visual recognition on open-vocabulary concepts. This is achieved by selecting the object category whose textual representation bears the highest similarity with the query image. While successful in some domains, this method struggles with identifying fine-grained entities as well as generalizing to unseen concepts that are not captured by the training distribution. Recent works attempt to mitigate these challenges by integrating category descriptions at test time, albeit yielding modest improvements. We attribute these limited gains to a fundamental misalignment between image and description representations, which is rooted in the pretraining structure of CLIP. In this paper, we propose GRAIN, a new pretraining strategy aimed at aligning representations at both fine and coarse levels simultaneously. Our approach learns to jointly ground textual descriptions in image regions along with aligning overarching captions with global image representations. To drive this pre-training, we leverage frozen Multimodal Large Language Models (MLLMs) to derive large-scale synthetic annotations. We demonstrate the enhanced zero-shot performance of our model compared to current state-of-the art methods across 11 diverse image classification datasets. Additionally, we introduce Products-2023, a newly curated, manually labeled dataset featuring novel concepts, and showcase our model's ability to recognize these concepts by benchmarking on it. Significant improvements achieved by our model on other downstream tasks like retrieval further highlight the superior quality of representations learned by our approach. Code available at https://github.com/shaunak27/grain-clip .
Community shame:
Not yet rated
Community Contributions

Found the code? Know the venue? Think something is wrong? Let us know!

๐Ÿ“œ Similar Papers

In the same crypt โ€” Computer Vision

๐ŸŒ… ๐ŸŒ… Old Age

Fast R-CNN

Ross Girshick

cs.CV ๐Ÿ› ICCV ๐Ÿ“š 27.7K cites 11 years ago