Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

DenseCap: Fully Convolutional Localization Networks for Dense Captioning

About

We introduce the dense captioning task, which requires a computer vision system to both localize and describe salient regions in images in natural language. The dense captioning task generalizes object detection when the descriptions consist of a single word, and Image Captioning when one predicted region covers the full image. To address the localization and description task jointly we propose a Fully Convolutional Localization Network (FCLN) architecture that processes an image with a single, efficient forward pass, requires no external regions proposals, and can be trained end-to-end with a single round of optimization. The architecture is composed of a Convolutional Network, a novel dense localization layer, and Recurrent Neural Network language model that generates the label sequences. We evaluate our network on the Visual Genome dataset, which comprises 94,000 images and 4,100,000 region-grounded captions. We observe both speed and accuracy improvements over baselines based on current state of the art approaches in both generation and retrieval settings.

Justin Johnson, Andrej Karpathy, Li Fei-Fei• 2015

Related benchmarks

TaskDatasetResultRank
Phrase DetectionVRD (test)
R@500.62
36
Dense CaptioningVG V1.0
mAP0.27
16
Dense CaptioningVisual Genome
mAP5.39
16
Paragraph GenerationStanford Image Paragraph Dataset 1.0 (test)
METEOR14.23
14
Dense CaptioningVisual Genome (test)
mAP27.03
13
Dense CaptioningVG V1.2
mAP5.2
13
Dense Object CaptioningVisual Genome 23
mAP5.39
9
Dense CaptioningVG 1.2 (test)
AP10
8
Relation RetrievalVRD (test)
Rr@54.16
7
Dense CaptioningVisual Genome Karpathy (test)
mAP5.39
7
Showing 10 of 21 rows

Other info

Code

Follow for update