Flickr30k Entities: Collecting Region-to-Phrase Correspondences for Richer Image-to-Sentence Models
By Bryan A. Plummer, Liwei Wang, Christopher M. Cervantes, Juan C. Caicedo, Julia Hockenmaier, Svetlana Lazebnik. 2015
2,641 citations recorded by the source corpus.
The Flickr30k dataset has become a standard benchmark for sentence-based image description. This paper presents Flickr30k Entities, which augments the 158k captions from Flickr30k with 244k coreference chains, linking mentions of the same entities across different captions for the same image, and associating them with 276k manually annotated bounding boxes. Such annotations are essential for continued progress in automatic image description and grounded language understanding. They enable us to define a new benchmark for localization of textual entity mentions in an image. We present a strong baseline for this task that combines an image-text embedding, detectors for common objects, a color classifier, and a bias towards selecting larger objects. While our baseline rivals in accuracy more complex state-of-the-art models, we show that its gains cannot be easily parlayed into improvements on such tasks as image-sentence retrieval, thus underlining the limitations of current methods and the need for further research.
Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Multimodal Learning
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
In Computer Vision and Pattern Recognition
- Luc Van Gool
- Dacheng Tao
- Trevor Darrell
- Xiaogang Wang
- Ming-Hsuan Yang
- Andrew Zisserman
- Jitendra Malik
- Cordelia Schmid
In Multimodal Learning
Citations in the corpus
- UNITER: Learning UNiversal Image-TExt Representations (2019, methodology): “Image-Text Retrieval Two datasets are considered for this task: COCO and Flickr30K.”
- VisualBERT: A Simple and Performant Baseline for Vision and Language (2019, methodology): “(2019)), (3) natural language for visual reasoning (NLVR 2 , Suhr et al. (2019)), and (4) region-to-phrase grounding (Flickr30K, Plummer et al. (2015)).”
- TAP: Text-Aware Pre-training for Text-VQA and Text-Caption (2020, methodology): “In general image-caption datasets [45, 29, 54, 41, 43], many image-text pairs contain either no scene text-related visual regions or no scene text-related language referring, and are thus less helpful to Text-VQA/TextCaption.”
- Self-Supervised MultiModal Versatile Networks (2020, background): “With larger datasets [48, 69, 71, 92, 96], many works have profited from learning such a joint visual-textual space [17, 19, 29, 30, 41, 53, 57, 63, 68, 84, 85, 87, 88, 93].”
- Large-Scale Adversarial Training for Vision-and-Language Representation Learning (2020, positive): “We consider five most common visual coreference types in Flickr30k Entities [52] and top five visual relations in Visual Genome [28] (listed in Table 4), and calculate the attention weights between region and phrase (or between regions) learned by pre-trained models.”
- Webly Supervised Joint Embedding for Cross-Modal Image-Text Retrieval (2018, methodology): “Flickr30K is another standard benchmark dataset for sentence-based image description.”
- BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation (2022, methodology): “We use the Karpathy split (Karpathy & Li, 2015) for both COCO and Flickr30K.”
- FLAVA: A Foundational Language And Vision Alignment Model (2021, methodology): “We also evaluate the FLAVA model on the multimodal zero-shot retrieval tasks over the Flickr30K and COCO datasets, where 2We follow hyperparameters used in FairSeq RoBERTa repo for finetuning on GLUE tasks without any further sweeping. the model needs to select a text caption based on a query image ”
- Uni-Perceiver: Pre-training Unified Architecture for Generic Perception for Zero-shot and Few-shot Tasks (2021, background): “, Flickr30k [60], MSVD [11], VQA [26], and GLUE [81].”
- Florence: A New Foundation Model for Computer Vision (2021, background): “Moreover, our fine tuning on retrieval is more efficient, with only roughly 6% and 8% fine-tuning epochs of ALIGN (Jia et al., 2021) on Flickr30k and MSCOCO respectively.”
- FILIP: Fine-grained Interactive Language-Image Pre-Training (2021, methodology): “We measure models’ performance on MSCOCO zero-7 Table 3: Results of fine-tuned image-text retrieval on Flickr30K and MSCOCO datasets.”
- VLMo: Unified Vision-Language Pre-Training with Mixture-of-Modality-Experts (2021, methodology): “We evaluate the model on the widely used COCO [27] and Flickr30K [33] datasets, and use the Karpathy split [19] for both datasets.”
- Align before Fuse: Vision and Language Representation Learning with Momentum Distillation (2021, methodology): “For zero-shot retrieval on Flickr30K, we evaluate with the model fine-tuned on COCO.”
- Conceptual 12M: Pushing Web-Scale Image-Text Pre-Training To Recognize Long-Tail Visual Concepts (2021, methodology): “The Flickr30K dataset [60] consists of 31,000 images from Flickr, each associated with five captions.”
- Scaling Up Visual and Vision-Language Representation Learning With Noisy Text Supervision (2021, methodology): “Image-text retrieval results on Flickr30K and MSCOCO datasets (zero-shot and fine-tuned).”
- VaTeX: A Large-Scale, High-Quality Multilingual Dataset for Video-and-Language Research (2019, methodology): “The Multi30K dataset [20], which is annotated based on the image captioning dataset Flickr30K [43], is commonly used in this direction.”
- Visual Dialog (2016, background): “A number of problems at the intersection of vision and language have recently gained prominence – image captioning [10, 11, 20, 54], video/movie description [43, 51, 52], text-to-image coreference/grounding [16, 22, 37, 39, 42], Visual Madlibs [60], and of course, visual question answering (VQA) [4,”