A Neural Network to Retrieve Images from Text Queries
نویسندگان
چکیده
This work presents a neural network for the retrieval of images from text queries. The proposed network is composed of two main modules: the first one extracts a global picture representation from local block descriptors while the second one aims at solving the retrieval problem from the extracted representation. Both modules are trained jointly to minimize a loss related to the retrieval performance. This approach is shown to be advantageous when compared to previous models relying on unsupervised feature extraction: average precision over Corel queries reaches 26.2% for our model, which should be compared to 21.6% for PAMIR, the best alternative.
منابع مشابه
Multimodal Neural Language Models
We introduce two multimodal neural language models: models of natural language that can be conditioned on other modalities. An image-text multimodal neural language model can be used to retrieve images given complex sentence queries, retrieve phrase descriptions given image queries, as well as generate text conditioned on images. We show that in the case of image-text modelling we can jointly l...
متن کاملLearning Document Image Features With SqueezeNet Convolutional Neural Network
The classification of various document images is considered an important step towards building a modern digital library or office automation system. Convolutional Neural Network (CNN) classifiers trained with backpropagation are considered to be the current state of the art model for this task. However, there are two major drawbacks for these classifiers: the huge computational power demand for...
متن کاملA Radon-based Convolutional Neural Network for Medical Image Retrieval
Image classification and retrieval systems have gained more attention because of easier access to high-tech medical imaging. However, the lack of availability of large-scaled balanced labelled data in medicine is still a challenge. Simplicity, practicality, efficiency, and effectiveness are the main targets in medical domain. To achieve these goals, Radon transformation, which is a well-known t...
متن کاملA Modified Grasshopper Optimization Algorithm Combined with CNN for Content Based Image Retrieval
Nowadays, with huge progress in digital imaging, new image processing methods are needed to manage digital images stored on disks. Image retrieval has been one of the most challengeable fields in digital image processing which means searching in a big database in order to represent similar images to the query image. Although many efficient researches have been performed for this topic so far, t...
متن کاملUsing Text Surrounding Method to Enhance Retrieval of Online Images by Google Search Engine
Purpose: the current research aimed to compare the effectiveness of various tags and codes for retrieving images from the Google. Design/methodology: selected images with different characteristics in a registered domain were carefully studied. The exception was that special conceptual features have been apportioned for each group of images separately. In this regard, each group image surr...
متن کامل